Also URL addition in upgrade

.github/workflows/pypi-release.yml

@@ -61,8 +61,8 @@ jobs:
         
         # --- API test ---
         # This also means that the docs/api-spec.yml was shipped and could be read
-        test -f /tmp/url-watches.json
-        API_KEY=$(jq -r '.. | .api_access_token? // empty' /tmp/url-watches.json)
+        test -f /tmp/changedetection.json
+        API_KEY=$(jq -r '.. | .api_access_token? // empty' /tmp/changedetection.json)
         echo Test API KEY is $API_KEY
         curl -X POST "http://127.0.0.1:10000/api/v1/watch" \
           -H "x-api-key: ${API_KEY}" \

.github/workflows/test-stack-reusable-workflow.yml

@@ -37,10 +37,29 @@ jobs:
             ${{ runner.os }}-pip-py${{ env.PYTHON_VERSION }}-
             ${{ runner.os }}-pip-
 
+      - name: Get current date for cache key
+        id: date
+        run: echo "date=$(date +'%Y-%m-%d')" >> $GITHUB_OUTPUT
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
       - name: Build changedetection.io container for testing under Python ${{ env.PYTHON_VERSION }}
+        uses: docker/build-push-action@v6
+        with:
+          context: ./
+          file: ./Dockerfile
+          build-args: |
+            PYTHON_VERSION=${{ env.PYTHON_VERSION }}
+            LOGGER_LEVEL=TRACE
+          tags: test-changedetectionio
+          load: true
+          cache-from: type=gha,scope=build-${{ github.ref_name }}-py${{ env.PYTHON_VERSION }}-${{ hashFiles('requirements.txt', 'Dockerfile') }}-${{ steps.date.outputs.date }}
+          cache-to: type=gha,mode=max,scope=build-${{ github.ref_name }}-py${{ env.PYTHON_VERSION }}-${{ hashFiles('requirements.txt', 'Dockerfile') }}-${{ steps.date.outputs.date }}
+
+      - name: Verify build
         run: |
-          echo "---- Building for Python ${{ env.PYTHON_VERSION }} -----"
-          docker build --build-arg PYTHON_VERSION=${{ env.PYTHON_VERSION }} --build-arg LOGGER_LEVEL=TRACE -t test-changedetectionio .
+          echo "---- Built for Python ${{ env.PYTHON_VERSION }} -----"
           docker run test-changedetectionio bash -c 'pip list'
 
       - name: We should be Python ${{ env.PYTHON_VERSION }} ...
@@ -84,7 +103,7 @@ jobs:
           docker run test-changedetectionio bash -c 'python3 -m unittest changedetectionio.tests.unit.test_watch_model'
           docker run test-changedetectionio bash -c 'python3 -m unittest changedetectionio.tests.unit.test_jinja2_security'
           docker run test-changedetectionio bash -c 'python3 -m unittest changedetectionio.tests.unit.test_semver'
-          docker run test-changedetectionio bash -c 'python3 -m unittest changedetectionio.tests.unit.test_html_to_text'
+          docker run test-changedetectionio bash -c 'python3 -m unittest changedetectionio.tests.unit.test_html_to_text'         
 
   # Basic pytest tests with ancillary services
   basic-tests:
@@ -111,6 +130,32 @@ jobs:
           docker network inspect changedet-network >/dev/null 2>&1 || docker network create changedet-network
           docker run --name test-cdio-basic-tests --network changedet-network test-changedetectionio bash -c 'cd changedetectionio && ./run_basic_tests.sh'
 
+      - name: Test CLI options
+        run: |
+          docker network inspect changedet-network >/dev/null 2>&1 || docker network create changedet-network
+          docker run --name test-cdio-cli-opts --network changedet-network test-changedetectionio bash -c 'changedetectionio/test_cli_opts.sh' &> cli-opts-output.txt
+          echo "=== CLI Options Test Output ==="
+          cat cli-opts-output.txt
+
+      - name: CLI Memory Test
+        run: |
+          echo "=== Checking CLI batch mode memory usage ==="
+          # Extract RSS memory value from output
+          RSS_MB=$(grep -oP "Memory consumption before worker shutdown: RSS=\K[\d.]+" cli-opts-output.txt | head -1 || echo "0")
+          echo "RSS Memory: ${RSS_MB} MB"
+
+          # Check if RSS is less than 100MB
+          if [ -n "$RSS_MB" ]; then
+            if (( $(echo "$RSS_MB < 100" | bc -l) )); then
+              echo "✓ Memory usage is acceptable: ${RSS_MB} MB < 100 MB"
+            else
+              echo "✗ Memory usage too high: ${RSS_MB} MB >= 100 MB"
+              exit 1
+            fi
+          else
+            echo "⚠ Could not extract memory usage, skipping check"
+          fi
+
       - name: Extract memory report and logs
         if: always()
         uses: ./.github/actions/extract-memory-report
@@ -125,6 +170,13 @@ jobs:
           name: test-cdio-basic-tests-output-py${{ env.PYTHON_VERSION }}
           path: output-logs
 
+      - name: Store CLI test output
+        if: always()
+        uses: actions/upload-artifact@v6
+        with:
+          name: test-cdio-cli-opts-output-py${{ env.PYTHON_VERSION }}
+          path: cli-opts-output.txt
+
   # Playwright tests
   playwright-tests:
     runs-on: ubuntu-latest
@@ -343,6 +395,29 @@ jobs:
           cd changedetectionio
           ./run_custom_browser_url_tests.sh
 
+  processor-plugin-tests:
+    runs-on: ubuntu-latest
+    needs: build
+    timeout-minutes: 20
+    env:
+      PYTHON_VERSION: ${{ inputs.python-version }}
+    steps:
+      - uses: actions/checkout@v6
+
+      - name: Download Docker image artifact
+        uses: actions/download-artifact@v7
+        with:
+          name: test-changedetectionio-${{ env.PYTHON_VERSION }}
+          path: /tmp
+
+      - name: Load Docker image
+        run: |
+          docker load -i /tmp/test-changedetectionio.tar
+
+      - name: Basic processor plugin registration and checks
+        run: |
+          docker run  -e EXTRA_PACKAGES=changedetection.io-osint-processor test-changedetectionio bash -c 'cd changedetectionio;pytest -vvv -s  tests/plugins/test_processor.py::test_check_plugin_processor'
+
   # Container startup tests
   container-tests:
     runs-on: ubuntu-latest
@@ -441,3 +516,142 @@ jobs:
             exit 1
           fi
           docker rm sig-test
+
+  # Upgrade path test
+  upgrade-path-test:
+    runs-on: ubuntu-latest
+    needs: build
+    timeout-minutes: 25
+    env:
+      PYTHON_VERSION: ${{ inputs.python-version }}
+    steps:
+      - uses: actions/checkout@v6
+        with:
+          fetch-depth: 0  # Fetch all history and tags for upgrade testing
+
+      - name: Set up Python ${{ env.PYTHON_VERSION }}
+        uses: actions/setup-python@v6
+        with:
+          python-version: ${{ env.PYTHON_VERSION }}
+
+      - name: Check upgrade works without error
+        run: |
+          echo "=== Testing upgrade path from 0.49.1 to ${{ github.ref_name }} (${{ github.sha }}) ==="
+
+          # Checkout old version and create datastore
+          git checkout 0.49.1
+          python3 -m venv .venv
+          source .venv/bin/activate
+          pip install -r requirements.txt
+          pip install 'pyOpenSSL>=23.2.0'
+
+          echo "=== Running version 0.49.1 to create datastore ==="
+          python3 ./changedetection.py -C -d /tmp/data &
+          APP_PID=$!
+
+          # Wait for app to be ready
+          echo "Waiting for 0.49.1 to be ready..."
+          sleep 6
+
+          # Extract API key from datastore (0.49.1 uses url-watches.json)
+          API_KEY=$(jq -r '.settings.application.api_access_token // empty' /tmp/data/url-watches.json)
+          echo "API Key: ${API_KEY:0:8}..."
+
+          # Create a watch with tag "github-group-test" via API
+          echo "Creating test watch with tag via API..."
+          curl -X POST "http://127.0.0.1:5000/api/v1/watch" \
+            -H "x-api-key: ${API_KEY}" \
+            -H "Content-Type: application/json" \
+            --show-error --fail \
+            --retry 6 --retry-delay 1 --retry-connrefused \
+            -d '{
+              "url": "https://example.com/upgrade-test",
+              "tag": "github-group-test"
+            }'
+
+          echo "✓ Created watch with tag 'github-group-test'"
+
+          # Create a specific test URL watch
+          echo "Creating test URL watch via API..."
+          curl -X POST "http://127.0.0.1:5000/api/v1/watch" \
+            -H "x-api-key: ${API_KEY}" \
+            -H "Content-Type: application/json" \
+            --show-error --fail \
+            -d '{
+              "url": "http://localhost/test.txt"
+            }'
+
+          echo "✓ Created watch for 'http://localhost/test.txt' in version 0.49.1"
+
+          # Stop the old version gracefully
+          kill $APP_PID
+          wait $APP_PID || true
+          echo "✓ Version 0.49.1 stopped"
+
+          # Upgrade to current version (use commit SHA since we're in detached HEAD)
+          echo "Upgrading to commit ${{ github.sha }}"
+          git checkout ${{ github.sha }}
+          pip install -r requirements.txt
+
+          echo "=== Running current version (commit ${{ github.sha }}) with old datastore (testing mode) ==="
+          TESTING_SHUTDOWN_AFTER_DATASTORE_LOAD=1 python3 ./changedetection.py -d /tmp/data > /tmp/upgrade-test.log 2>&1
+
+          echo "=== Upgrade test output ==="
+          cat /tmp/upgrade-test.log
+          echo "✓ Datastore upgraded successfully"
+
+          # Now start the current version normally to verify the tag survived
+          echo "=== Starting current version to verify tag exists after upgrade ==="
+          timeout 20 python3 ./changedetection.py -d /tmp/data > /tmp/ui-test.log 2>&1 &
+          APP_PID=$!
+
+          # Wait for app to be ready and fetch UI
+          echo "Waiting for current version to be ready..."
+          sleep 5
+          curl --retry 6 --retry-delay 1 --retry-connrefused --silent http://127.0.0.1:5000 > /tmp/ui-output.html
+
+          # Verify tag exists in UI
+          if grep -q "github-group-test" /tmp/ui-output.html; then
+            echo "✓ Tag 'github-group-test' found in UI after upgrade"
+          else
+            echo "ERROR: Tag 'github-group-test' not found in UI after upgrade"
+            echo "=== UI Output ==="
+            cat /tmp/ui-output.html
+            echo "=== App Log ==="
+            cat /tmp/ui-test.log
+            kill $APP_PID || true
+            exit 1
+          fi
+
+          # Verify test URL exists in UI
+          if grep -q "http://localhost/test.txt" /tmp/ui-output.html; then
+            echo "✓ Watch URL 'http://localhost/test.txt' found in UI after upgrade"
+          else
+            echo "ERROR: Watch URL 'http://localhost/test.txt' not found in UI after upgrade"
+            echo "=== UI Output ==="
+            cat /tmp/ui-output.html
+            echo "=== App Log ==="
+            cat /tmp/ui-test.log
+            kill $APP_PID || true
+            exit 1
+          fi
+
+          # Cleanup
+          kill $APP_PID || true
+          wait $APP_PID || true
+
+          echo ""
+          echo "✓✓✓ Upgrade test passed: 0.49.1 → ${{ github.ref_name }} ✓✓✓"
+          echo "    - Commit: ${{ github.sha }}"
+          echo "    - Datastore migrated successfully"
+          echo "    - Tag 'github-group-test' survived upgrade"
+          echo "    - Watch URL 'http://localhost/test.txt' survived upgrade"
+
+          echo "✓ Upgrade test passed: 0.49.1 → ${{ github.ref_name }}"
+
+      - name: Upload upgrade test logs
+        if: always()
+        uses: actions/upload-artifact@v6
+        with:
+          name: upgrade-test-logs-py${{ env.PYTHON_VERSION }}
+          path: /tmp/upgrade-test.log

.gitignore

@@ -29,3 +29,4 @@ test-datastore/
 
 # Memory consumption log
 test-memory.log
+tests/logs/

Dockerfile

@@ -138,6 +138,15 @@ ENV LOGGER_LEVEL="$LOGGER_LEVEL"
 ENV LC_ALL=en_US.UTF-8
 
 WORKDIR /app
+
+# Copy and set up entrypoint script for installing extra packages
+COPY docker-entrypoint.sh /docker-entrypoint.sh
+RUN chmod +x /docker-entrypoint.sh
+
+# Set entrypoint to handle EXTRA_PACKAGES env var
+ENTRYPOINT ["/docker-entrypoint.sh"]
+
+# Default command (can be overridden in docker-compose.yml)
 CMD ["python", "./changedetection.py", "-d", "/datastore"]
 
 

MANIFEST.in

@@ -9,6 +9,7 @@ recursive-include changedetectionio/notification *
 recursive-include changedetectionio/processors *
 recursive-include changedetectionio/realtime *
 recursive-include changedetectionio/static *
+recursive-include changedetectionio/store *
 recursive-include changedetectionio/templates *
 recursive-include changedetectionio/tests *
 recursive-include changedetectionio/translations *

changedetectionio/__init__.py

@@ -2,23 +2,24 @@
 
 # Read more https://github.com/dgtlmoon/changedetection.io/wiki
 # Semver means never use .01, or 00. Should be .1.
-__version__ = '0.52.8'
+__version__ = '0.52.9'
 
 from changedetectionio.strtobool import strtobool
 from json.decoder import JSONDecodeError
+
+from loguru import logger
+import getopt
 import logging
 import os
-import getopt
 import platform
 import signal
-
-import sys
+import threading
+import time
 
 # Eventlet completely removed - using threading mode for SocketIO
 # This provides better Python 3.12+ compatibility and eliminates eventlet/asyncio conflicts
-from changedetectionio import store
-from changedetectionio.flask_app import changedetection_app
-from loguru import logger
+# Note: store and changedetection_app are imported inside main() to avoid
+# initialization before argument parsing (allows --help to work without loading everything)
 
 # ==============================================================================
 # Multiprocessing Configuration - CRITICAL for Thread Safety
@@ -83,15 +84,26 @@ def get_version():
 def sigshutdown_handler(_signo, _stack_frame):
     name = signal.Signals(_signo).name
     logger.critical(f'Shutdown: Got Signal - {name} ({_signo}), Fast shutdown initiated')
-    
+
     # Set exit flag immediately to stop all loops
     app.config.exit.set()
     datastore.stop_thread = True
-    
+
+    # Log memory consumption before shutting down workers (cross-platform)
+    try:
+        import psutil
+        process = psutil.Process()
+        mem_info = process.memory_info()
+        rss_mb = mem_info.rss / 1024 / 1024
+        vms_mb = mem_info.vms / 1024 / 1024
+        logger.info(f"Memory consumption before worker shutdown: RSS={rss_mb:,.2f} MB, VMS={vms_mb:,.2f} MB")
+    except Exception as e:
+        logger.warning(f"Could not retrieve memory stats: {str(e)}")
+
     # Shutdown workers and queues immediately
     try:
-        from changedetectionio import worker_handler
-        worker_handler.shutdown_workers()
+        from changedetectionio import worker_pool
+        worker_pool.shutdown_workers()
     except Exception as e:
         logger.error(f"Error shutting down workers: {str(e)}")
     
@@ -100,9 +112,9 @@ def sigshutdown_handler(_signo, _stack_frame):
         from changedetectionio.flask_app import update_q, notification_q
         update_q.close()
         notification_q.close()
-        logger.debug("Janus queues closed successfully")
+        logger.debug("Queues closed successfully")
     except Exception as e:
-        logger.critical(f"CRITICAL: Failed to close janus queues: {e}")
+        logger.critical(f"CRITICAL: Failed to close queues: {e}")
     
     # Shutdown socketio server fast
     from changedetectionio.flask_app import socketio_server
@@ -112,31 +124,80 @@ def sigshutdown_handler(_signo, _stack_frame):
         except Exception as e:
             logger.error(f"Error shutting down Socket.IO server: {str(e)}")
     
-    # Save data quickly
-    try:
-        datastore.sync_to_json()
-        logger.success('Fast sync to disk complete.')
-    except Exception as e:
-        logger.error(f"Error syncing to disk: {str(e)}")
-    
+    # With immediate persistence, all data is already saved
+    logger.success('All data already persisted (immediate commits enabled).')
+
     sys.exit()
 
+def print_help():
+    """Print help text for command line options"""
+    print('Usage: changedetection.py [options]')
+    print('')
+    print('Standard options:')
+    print('  -s                SSL enable')
+    print('  -h HOST           Listen host (default: 0.0.0.0)')
+    print('  -p PORT           Listen port (default: 5000)')
+    print('  -d PATH           Datastore path')
+    print('  -l LEVEL          Log level (TRACE, DEBUG, INFO, SUCCESS, WARNING, ERROR, CRITICAL)')
+    print('  -c                Cleanup unused snapshots')
+    print('  -C                Create datastore directory if it doesn\'t exist')
+    print('  -P true/false     Set all watches paused (true) or active (false)')
+    print('')
+    print('Add URLs on startup:')
+    print('  -u URL            Add URL to watch (can be used multiple times)')
+    print('  -u0 \'JSON\'        Set options for first -u URL (e.g. \'{"processor":"text_json_diff"}\')')
+    print('  -u1 \'JSON\'        Set options for second -u URL (0-indexed)')
+    print('  -u2 \'JSON\'        Set options for third -u URL, etc.')
+    print('                    Available options: processor, fetch_backend, headers, method, etc.')
+    print('                    See model/Watch.py for all available options')
+    print('')
+    print('Recheck on startup:')
+    print('  -r all            Queue all watches for recheck on startup')
+    print('  -r UUID,...       Queue specific watches (comma-separated UUIDs)')
+    print('  -r all N          Queue all watches, wait for completion, repeat N times')
+    print('  -r UUID,... N     Queue specific watches, wait for completion, repeat N times')
+    print('')
+    print('Batch mode:')
+    print('  -b                Run in batch mode (process queue then exit)')
+    print('                    Useful for CI/CD, cron jobs, or one-time checks')
+    print('                    NOTE: Batch mode checks if Flask is running and aborts if port is in use')
+    print('                    Use -p PORT to specify a different port if needed')
+    print('')
+
 def main():
     global datastore
     global app
 
+    # Early help/version check before any initialization
+    if '--help' in sys.argv or '-help' in sys.argv:
+        print_help()
+        sys.exit(0)
+
+    if '--version' in sys.argv or '-v' in sys.argv:
+        print(f'changedetection.io {__version__}')
+        sys.exit(0)
+
+    # Import heavy modules after help/version checks to keep startup fast for those flags
+    from changedetectionio import store
+    from changedetectionio.flask_app import changedetection_app
+
     datastore_path = None
-    do_cleanup = False
-    # Optional URL to watch since start
-    default_url = None
     # Set a default logger level
     logger_level = 'DEBUG'
     include_default_watches = True
+    all_paused = None  # None means don't change, True/False to set
 
     host = os.environ.get("LISTEN_HOST", "0.0.0.0").strip()
     port = int(os.environ.get('PORT', 5000))
     ssl_mode = False
 
+    # Lists for multiple URLs and their options
+    urls_to_add = []
+    url_options = {}  # Key: index (0-based), Value: dict of options
+    recheck_watches = None  # None, 'all', or list of UUIDs
+    recheck_repeat_count = 1  # Number of times to repeat recheck cycle
+    batch_mode = False  # Run once then exit when queue is empty
+
     # On Windows, create and use a default path.
     if os.name == 'nt':
         datastore_path = os.path.expandvars(r'%APPDATA%\changedetection.io')
@@ -145,10 +206,68 @@ def main():
         # Must be absolute so that send_from_directory doesnt try to make it relative to backend/
         datastore_path = os.path.join(os.getcwd(), "../datastore")
 
+    # Pre-process arguments to extract -u, -u<N>, and -r options before getopt
+    # This allows unlimited -u0, -u1, -u2, ... options without predefining them
+    cleaned_argv = ['changedetection.py']  # Start with program name
+    i = 1
+    while i < len(sys.argv):
+        arg = sys.argv[i]
+
+        # Handle -u (add URL)
+        if arg == '-u' and i + 1 < len(sys.argv):
+            urls_to_add.append(sys.argv[i + 1])
+            i += 2
+            continue
+
+        # Handle -u<N> (set options for URL at index N)
+        if arg.startswith('-u') and len(arg) > 2 and arg[2:].isdigit():
+            idx = int(arg[2:])
+            if i + 1 < len(sys.argv):
+                try:
+                    import json
+                    url_options[idx] = json.loads(sys.argv[i + 1])
+                except json.JSONDecodeError as e:
+                    print(f'Error: Invalid JSON for {arg}: {sys.argv[i + 1]}')
+                    print(f'JSON decode error: {e}')
+                    sys.exit(2)
+                i += 2
+                continue
+
+        # Handle -r (recheck watches)
+        if arg == '-r' and i + 1 < len(sys.argv):
+            recheck_arg = sys.argv[i + 1]
+            if recheck_arg.lower() == 'all':
+                recheck_watches = 'all'
+            else:
+                # Parse comma-separated list of UUIDs
+                recheck_watches = [uuid.strip() for uuid in recheck_arg.split(',') if uuid.strip()]
+
+            # Check for optional repeat count as third argument
+            if i + 2 < len(sys.argv) and sys.argv[i + 2].isdigit():
+                recheck_repeat_count = int(sys.argv[i + 2])
+                if recheck_repeat_count < 1:
+                    print(f'Error: Repeat count must be at least 1, got {recheck_repeat_count}')
+                    sys.exit(2)
+                i += 3
+            else:
+                i += 2
+            continue
+
+        # Handle -b (batch mode - run once and exit)
+        if arg == '-b':
+            batch_mode = True
+            i += 1
+            continue
+
+        # Keep other arguments for getopt
+        cleaned_argv.append(arg)
+        i += 1
+
     try:
-        opts, args = getopt.getopt(sys.argv[1:], "6Ccsd:h:p:l:u:", "port")
-    except getopt.GetoptError:
-        print('backend.py -s SSL enable -h [host] -p [port] -d [datastore path] -u [default URL to watch] -l [debug level - TRACE, DEBUG(default), INFO, SUCCESS, WARNING, ERROR, CRITICAL]')
+        opts, args = getopt.getopt(cleaned_argv[1:], "6Csd:h:p:l:P:", "port")
+    except getopt.GetoptError as e:
+        print_help()
+        print(f'Error: {e}')
         sys.exit(2)
 
     create_datastore_dir = False
@@ -173,14 +292,6 @@ def main():
         if opt == '-d':
             datastore_path = arg
 
-        if opt == '-u':
-            default_url = arg
-            include_default_watches = False
-
-        # Cleanup (remove text files that arent in the index)
-        if opt == '-c':
-            do_cleanup = True
-
         # Create the datadir if it doesnt exist
         if opt == '-C':
             create_datastore_dir = True
@@ -188,6 +299,18 @@ def main():
         if opt == '-l':
             logger_level = int(arg) if arg.isdigit() else arg.upper()
 
+        if opt == '-P':
+            try:
+                all_paused = bool(strtobool(arg))
+            except ValueError:
+                print(f'Error: Invalid value for -P option: {arg}')
+                print('Expected: true, false, yes, no, 1, or 0')
+                sys.exit(2)
+
+    # If URLs are provided, don't include default watches
+    if urls_to_add:
+        include_default_watches = False
+
 
     logger.success(f"changedetection.io version {get_version()} starting.")
     # Launch using SocketIO run method for proper integration (if enabled)
@@ -224,11 +347,16 @@ def main():
     logging.getLogger('pyppeteer.connection.Connection').setLevel(logging.WARNING)
 
     # isnt there some @thingy to attach to each route to tell it, that this route needs a datastore
-    app_config = {'datastore_path': datastore_path}
+    app_config = {
+        'datastore_path': datastore_path,
+        'batch_mode': batch_mode,
+        'recheck_watches': recheck_watches,
+        'recheck_repeat_count': recheck_repeat_count
+    }
 
     if not os.path.isdir(app_config['datastore_path']):
         if create_datastore_dir:
-            os.mkdir(app_config['datastore_path'])
+            os.makedirs(app_config['datastore_path'], exist_ok=True)
         else:
             logger.critical(
                 f"ERROR: Directory path for the datastore '{app_config['datastore_path']}'"
@@ -243,17 +371,219 @@ def main():
         # Dont' start if the JSON DB looks corrupt
         logger.critical(f"ERROR: JSON DB or Proxy List JSON at '{app_config['datastore_path']}' appears to be corrupt, aborting.")
         logger.critical(str(e))
-        return
+        sys.exit(1)
+
+    # Testing mode: Exit cleanly after datastore initialization (for CI/CD upgrade tests)
+    if os.environ.get('TESTING_SHUTDOWN_AFTER_DATASTORE_LOAD'):
+        logger.success(f"TESTING MODE: Datastore loaded successfully from {app_config['datastore_path']}")
+        logger.success(f"TESTING MODE: Schema version: {datastore.data['settings']['application'].get('schema_version', 'unknown')}")
+        logger.success(f"TESTING MODE: Loaded {len(datastore.data['watching'])} watches")
+        logger.success("TESTING MODE: Exiting cleanly (TESTING_SHUTDOWN_AFTER_DATASTORE_LOAD is set)")
+        sys.exit(0)
+
+    # Apply all_paused setting if specified via CLI
+    if all_paused is not None:
+        datastore.data['settings']['application']['all_paused'] = all_paused
+        logger.info(f"Setting all watches paused: {all_paused}")
 
     # Inject datastore into plugins that need access to settings
     from changedetectionio.pluggy_interface import inject_datastore_into_plugins
     inject_datastore_into_plugins(datastore)
 
-    if default_url:
-        datastore.add_watch(url = default_url)
+    # Step 1: Add URLs with their options (if provided via -u flags)
+    added_watch_uuids = []
+    if urls_to_add:
+        logger.info(f"Adding {len(urls_to_add)} URL(s) from command line")
+        for idx, url in enumerate(urls_to_add):
+            extras = url_options.get(idx, {})
+            if extras:
+                logger.debug(f"Adding watch {idx}: {url} with options: {extras}")
+            else:
+                logger.debug(f"Adding watch {idx}: {url}")
+
+            new_uuid = datastore.add_watch(url=url, extras=extras)
+            if new_uuid:
+                added_watch_uuids.append(new_uuid)
+                logger.success(f"Added watch: {url} (UUID: {new_uuid})")
+            else:
+                logger.error(f"Failed to add watch: {url}")
 
     app = changedetection_app(app_config, datastore)
 
+    # Step 2: Queue newly added watches (if -u was provided in batch mode)
+    # This must happen AFTER app initialization so update_q is available
+    if batch_mode and added_watch_uuids:
+        from changedetectionio.flask_app import update_q
+        from changedetectionio import queuedWatchMetaData, worker_pool
+
+        logger.info(f"Batch mode: Queuing {len(added_watch_uuids)} newly added watches")
+        for watch_uuid in added_watch_uuids:
+            try:
+                worker_pool.queue_item_async_safe(
+                    update_q,
+                    queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid})
+                )
+                logger.debug(f"Queued newly added watch: {watch_uuid}")
+            except Exception as e:
+                logger.error(f"Failed to queue watch {watch_uuid}: {e}")
+
+    # Step 3: Queue watches for recheck (if -r was provided)
+    # This must happen AFTER app initialization so update_q is available
+    if recheck_watches is not None:
+        from changedetectionio.flask_app import update_q
+        from changedetectionio import queuedWatchMetaData, worker_pool
+
+        watches_to_queue = []
+        if recheck_watches == 'all':
+            # Queue all watches, excluding those already queued in batch mode
+            all_watches = list(datastore.data['watching'].keys())
+            if batch_mode and added_watch_uuids:
+                # Exclude newly added watches that were already queued in batch mode
+                watches_to_queue = [uuid for uuid in all_watches if uuid not in added_watch_uuids]
+                logger.info(f"Queuing {len(watches_to_queue)} existing watches for recheck ({len(added_watch_uuids)} newly added watches already queued)")
+            else:
+                watches_to_queue = all_watches
+                logger.info(f"Queuing all {len(watches_to_queue)} watches for recheck")
+        else:
+            # Queue specific UUIDs
+            watches_to_queue = recheck_watches
+            logger.info(f"Queuing {len(watches_to_queue)} specific watches for recheck")
+
+        queued_count = 0
+        for watch_uuid in watches_to_queue:
+            if watch_uuid in datastore.data['watching']:
+                try:
+                    worker_pool.queue_item_async_safe(
+                        update_q,
+                        queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid})
+                    )
+                    queued_count += 1
+                    logger.debug(f"Queued watch for recheck: {watch_uuid}")
+                except Exception as e:
+                    logger.error(f"Failed to queue watch {watch_uuid}: {e}")
+            else:
+                logger.warning(f"Watch UUID not found in datastore: {watch_uuid}")
+
+        logger.success(f"Successfully queued {queued_count} watches for recheck")
+
+    # Step 4: Setup batch mode monitor (if -b was provided)
+    if batch_mode:
+        from changedetectionio.flask_app import update_q
+
+        # Safety check: Ensure Flask app is not already running on this port
+        # Batch mode should never run alongside the web server
+        import socket
+        test_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+
+        try:
+            # Try to bind to the configured host:port (no SO_REUSEADDR - strict check)
+            test_socket.bind((host, port))
+            test_socket.close()
+            logger.debug(f"Batch mode: Port {port} is available (Flask app not running)")
+        except OSError as e:
+            test_socket.close()
+            # errno 98 = EADDRINUSE (Linux)
+            # errno 48 = EADDRINUSE (macOS)
+            # errno 10048 = WSAEADDRINUSE (Windows)
+            if e.errno in (48, 98, 10048) or "Address already in use" in str(e) or "already in use" in str(e).lower():
+                logger.critical(f"ERROR: Batch mode cannot run - port {port} is already in use")
+                logger.critical(f"The Flask web server appears to be running on {host}:{port}")
+                logger.critical(f"Batch mode is designed for standalone operation (CI/CD, cron jobs, etc.)")
+                logger.critical(f"Please either stop the Flask web server, or use a different port with -p PORT")
+                sys.exit(1)
+            else:
+                # Some other socket error - log but continue (might be network configuration issue)
+                logger.warning(f"Port availability check failed with unexpected error: {e}")
+                logger.warning(f"Continuing with batch mode anyway - be aware of potential conflicts")
+
+        def queue_watches_for_recheck(datastore, iteration):
+            """Helper function to queue watches for recheck"""
+            watches_to_queue = []
+            if recheck_watches == 'all':
+                all_watches = list(datastore.data['watching'].keys())
+                if batch_mode and added_watch_uuids and iteration == 1:
+                    # Only exclude newly added watches on first iteration
+                    watches_to_queue = [uuid for uuid in all_watches if uuid not in added_watch_uuids]
+                else:
+                    watches_to_queue = all_watches
+                logger.info(f"Batch mode (iteration {iteration}): Queuing all {len(watches_to_queue)} watches")
+            elif recheck_watches:
+                watches_to_queue = recheck_watches
+                logger.info(f"Batch mode (iteration {iteration}): Queuing {len(watches_to_queue)} specific watches")
+
+            queued_count = 0
+            for watch_uuid in watches_to_queue:
+                if watch_uuid in datastore.data['watching']:
+                    try:
+                        worker_pool.queue_item_async_safe(
+                            update_q,
+                            queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid})
+                        )
+                        queued_count += 1
+                    except Exception as e:
+                        logger.error(f"Failed to queue watch {watch_uuid}: {e}")
+                else:
+                    logger.warning(f"Watch UUID not found in datastore: {watch_uuid}")
+            logger.success(f"Batch mode (iteration {iteration}): Successfully queued {queued_count} watches")
+            return queued_count
+
+        def batch_mode_monitor():
+            """Monitor queue and workers, shutdown or repeat when work is complete"""
+            import time
+
+            # Track iterations if repeat mode is enabled
+            current_iteration = 1
+            total_iterations = recheck_repeat_count if recheck_watches and recheck_repeat_count > 1 else 1
+
+            if total_iterations > 1:
+                logger.info(f"Batch mode: Will repeat recheck {total_iterations} times")
+            else:
+                logger.info("Batch mode: Waiting for all queued items to complete...")
+
+            # Wait a bit for workers to start processing
+            time.sleep(3)
+
+            try:
+                while current_iteration <= total_iterations:
+                    logger.info(f"Batch mode: Waiting for iteration {current_iteration}/{total_iterations} to complete...")
+
+                    # Use the shared wait_for_all_checks function
+                    completed = worker_pool.wait_for_all_checks(update_q, timeout=300)
+
+                    if not completed:
+                        logger.warning(f"Batch mode: Iteration {current_iteration} timed out after 300 seconds")
+
+                    logger.success(f"Batch mode: Iteration {current_iteration}/{total_iterations} completed")
+
+                    # Check if we need to repeat
+                    if current_iteration < total_iterations:
+                        logger.info(f"Batch mode: Starting iteration {current_iteration + 1}...")
+                        current_iteration += 1
+
+                        # Re-queue watches for next iteration
+                        queue_watches_for_recheck(datastore, current_iteration)
+
+                        # Brief pause before continuing
+                        time.sleep(2)
+                    else:
+                        # All iterations complete
+                        logger.success(f"Batch mode: All {total_iterations} iterations completed, initiating shutdown")
+                        # Trigger shutdown
+                        import os, signal
+                        os.kill(os.getpid(), signal.SIGTERM)
+                        return
+
+            except Exception as e:
+                logger.error(f"Batch mode monitor error: {e}")
+                logger.error(f"Initiating emergency shutdown")
+                import os, signal
+                os.kill(os.getpid(), signal.SIGTERM)
+
+        # Start monitor in background thread
+        monitor_thread = threading.Thread(target=batch_mode_monitor, daemon=True, name="BatchModeMonitor")
+        monitor_thread.start()
+        logger.info("Batch mode enabled: Will exit after all queued items are processed")
+
     # Get the SocketIO instance from the Flask app (created in flask_app.py)
     from changedetectionio.flask_app import socketio_server
     global socketio
@@ -275,10 +605,6 @@ def main():
     else:
         logger.info("SIGUSR1 handler only registered on Linux, skipped.")
 
-    # Go into cleanup mode
-    if do_cleanup:
-        datastore.remove_unused_snapshots()
-
     app.config['datastore_path'] = datastore_path
 
 
@@ -287,7 +613,7 @@ def main():
         return dict(right_sticky="v{}".format(datastore.data['version_tag']),
                     new_version_available=app.config['NEW_VERSION_AVAILABLE'],
                     has_password=datastore.data['settings']['application']['password'] != False,
-                    socket_io_enabled=datastore.data['settings']['application']['ui'].get('socket_io_enabled', True),
+                    socket_io_enabled=datastore.data['settings']['application'].get('ui', {}).get('socket_io_enabled', True),
                     all_paused=datastore.data['settings']['application'].get('all_paused', False),
                     all_muted=datastore.data['settings']['application'].get('all_muted', False)
                     )
@@ -311,23 +637,43 @@ def main():
     if os.getenv('USE_X_SETTINGS'):
         logger.info("USE_X_SETTINGS is ENABLED")
         from werkzeug.middleware.proxy_fix import ProxyFix
-        app.wsgi_app = ProxyFix(app.wsgi_app, x_prefix=1, x_host=1)
+        app.wsgi_app = ProxyFix(
+            app.wsgi_app,
+            x_for=1,      # X-Forwarded-For (client IP)
+            x_proto=1,    # X-Forwarded-Proto (http/https)
+            x_host=1,     # X-Forwarded-Host (original host)
+            x_port=1,     # X-Forwarded-Port (original port)
+            x_prefix=1    # X-Forwarded-Prefix (URL prefix)
+        )
 
 
-    # SocketIO instance is already initialized in flask_app.py
-    if socketio_server:
-        if ssl_mode:
-            logger.success(f"SSL mode enabled, attempting to start with '{ssl_cert_file}' and '{ssl_privkey_file}' in {os.getcwd()}")
-            socketio.run(app, host=host, port=int(port), debug=False,
-                         ssl_context=(ssl_cert_file, ssl_privkey_file), allow_unsafe_werkzeug=True)
-        else:
-            socketio.run(app, host=host, port=int(port), debug=False, allow_unsafe_werkzeug=True)
+    # In batch mode, skip starting the HTTP server - just keep workers running
+    if batch_mode:
+        logger.info("Batch mode: Skipping HTTP server startup, workers will process queue")
+        logger.info("Batch mode: Main thread will wait for shutdown signal")
+        # Keep main thread alive until batch monitor triggers shutdown
+        try:
+            while True:
+                time.sleep(1)
+        except KeyboardInterrupt:
+            logger.info("Batch mode: Keyboard interrupt received")
+            pass
     else:
-        # Run Flask app without Socket.IO if disabled
-        logger.info("Starting Flask app without Socket.IO server")
-        if ssl_mode:
-            logger.success(f"SSL mode enabled, attempting to start with '{ssl_cert_file}' and '{ssl_privkey_file}' in {os.getcwd()}")
-            app.run(host=host, port=int(port), debug=False,
-                    ssl_context=(ssl_cert_file, ssl_privkey_file))
+        # Normal mode: Start HTTP server
+        # SocketIO instance is already initialized in flask_app.py
+        if socketio_server:
+            if ssl_mode:
+                logger.success(f"SSL mode enabled, attempting to start with '{ssl_cert_file}' and '{ssl_privkey_file}' in {os.getcwd()}")
+                socketio.run(app, host=host, port=int(port), debug=False,
+                             ssl_context=(ssl_cert_file, ssl_privkey_file), allow_unsafe_werkzeug=True)
+            else:
+                socketio.run(app, host=host, port=int(port), debug=False, allow_unsafe_werkzeug=True)
         else:
-            app.run(host=host, port=int(port), debug=False)
+            # Run Flask app without Socket.IO if disabled
+            logger.info("Starting Flask app without Socket.IO server")
+            if ssl_mode:
+                logger.success(f"SSL mode enabled, attempting to start with '{ssl_cert_file}' and '{ssl_privkey_file}' in {os.getcwd()}")
+                app.run(host=host, port=int(port), debug=False,
+                        ssl_context=(ssl_cert_file, ssl_privkey_file))
+            else:
+                app.run(host=host, port=int(port), debug=False)

changedetectionio/api/Import.py

@@ -2,8 +2,12 @@ from changedetectionio.strtobool import strtobool
 from flask_restful import abort, Resource
 from flask import request
 from functools import wraps
-from . import auth, validate_openapi_request
+from . import auth, validate_openapi_request, schema_create_watch
 from ..validate_url import is_safe_valid_url
+import json
+
+# Number of URLs above which import switches to background processing
+IMPORT_SWITCH_TO_BACKGROUND_THRESHOLD = 20
 
 
 def default_content_type(content_type='text/plain'):
@@ -19,6 +23,62 @@ def default_content_type(content_type='text/plain'):
     return decorator
 
 
+def convert_query_param_to_type(value, schema_property):
+    """
+    Convert a query parameter string to the appropriate type based on schema definition.
+
+    Args:
+        value: String value from query parameter
+        schema_property: Schema property definition with 'type' or 'anyOf' field
+
+    Returns:
+        Converted value in the appropriate type
+    """
+    # Handle anyOf schemas (extract the first type)
+    if 'anyOf' in schema_property:
+        # Use the first non-null type from anyOf
+        for option in schema_property['anyOf']:
+            if option.get('type') and option.get('type') != 'null':
+                prop_type = option.get('type')
+                break
+        else:
+            prop_type = None
+    else:
+        prop_type = schema_property.get('type')
+
+    # Handle array type (e.g., notification_urls)
+    if prop_type == 'array':
+        # Support both comma-separated and JSON array format
+        if value.startswith('['):
+            try:
+                return json.loads(value)
+            except json.JSONDecodeError:
+                return [v.strip() for v in value.split(',')]
+        return [v.strip() for v in value.split(',')]
+
+    # Handle object type (e.g., time_between_check, headers)
+    elif prop_type == 'object':
+        try:
+            return json.loads(value)
+        except json.JSONDecodeError:
+            raise ValueError(f"Invalid JSON object for field: {value}")
+
+    # Handle boolean type
+    elif prop_type == 'boolean':
+        return strtobool(value)
+
+    # Handle integer type
+    elif prop_type == 'integer':
+        return int(value)
+
+    # Handle number type (float)
+    elif prop_type == 'number':
+        return float(value)
+
+    # Default: return as string
+    return value
+
+
 class Import(Resource):
     def __init__(self, **kwargs):
         # datastore is a black box dependency
@@ -28,40 +88,127 @@ class Import(Resource):
     @default_content_type('text/plain') #3547 #3542
     @validate_openapi_request('importWatches')
     def post(self):
-        """Import a list of watched URLs."""
+        """Import a list of watched URLs with optional watch configuration."""
+
+        # Special parameters that are NOT watch configuration
+        special_params = {'tag', 'tag_uuids', 'dedupe', 'proxy'}
 
         extras = {}
 
+        # Handle special 'proxy' parameter
         if request.args.get('proxy'):
             plist = self.datastore.proxy_list
             if not request.args.get('proxy') in plist:
-                return "Invalid proxy choice, currently supported proxies are '{}'".format(', '.join(plist)), 400
+                proxy_list_str = ', '.join(plist) if plist else 'none configured'
+                return f"Invalid proxy choice, currently supported proxies are '{proxy_list_str}'", 400
             else:
                 extras['proxy'] = request.args.get('proxy')
 
+        # Handle special 'dedupe' parameter
         dedupe = strtobool(request.args.get('dedupe', 'true'))
 
+        # Handle special 'tag' and 'tag_uuids' parameters
         tags = request.args.get('tag')
         tag_uuids = request.args.get('tag_uuids')
 
         if tag_uuids:
             tag_uuids = tag_uuids.split(',')
 
+        # Extract ALL other query parameters as watch configuration
+        schema_properties = schema_create_watch.get('properties', {})
+        for param_name, param_value in request.args.items():
+            # Skip special parameters
+            if param_name in special_params:
+                continue
+
+            # Skip if not in schema (unknown parameter)
+            if param_name not in schema_properties:
+                return f"Unknown watch configuration parameter: {param_name}", 400
+
+            # Convert to appropriate type based on schema
+            try:
+                converted_value = convert_query_param_to_type(param_value, schema_properties[param_name])
+                extras[param_name] = converted_value
+            except (ValueError, json.JSONDecodeError) as e:
+                return f"Invalid value for parameter '{param_name}': {str(e)}", 400
+
+        # Validate processor if provided
+        if 'processor' in extras:
+            from changedetectionio.processors import available_processors
+            available = [p[0] for p in available_processors()]
+            if extras['processor'] not in available:
+                return f"Invalid processor '{extras['processor']}'. Available processors: {', '.join(available)}", 400
+
+        # Validate fetch_backend if provided
+        if 'fetch_backend' in extras:
+            from changedetectionio.content_fetchers import available_fetchers
+            available = [f[0] for f in available_fetchers()]
+            # Also allow 'system' and extra_browser_* patterns
+            is_valid = (
+                extras['fetch_backend'] == 'system' or
+                extras['fetch_backend'] in available or
+                extras['fetch_backend'].startswith('extra_browser_')
+            )
+            if not is_valid:
+                return f"Invalid fetch_backend '{extras['fetch_backend']}'. Available: system, {', '.join(available)}", 400
+
+        # Validate notification_urls if provided
+        if 'notification_urls' in extras:
+            from wtforms import ValidationError
+            from changedetectionio.api.Notifications import validate_notification_urls
+            try:
+                validate_notification_urls(extras['notification_urls'])
+            except ValidationError as e:
+                return f"Invalid notification_urls: {str(e)}", 400
+
         urls = request.get_data().decode('utf8').splitlines()
-        added = []
+        # Clean and validate URLs upfront
+        urls_to_import = []
         for url in urls:
             url = url.strip()
             if not len(url):
                 continue
 
-            # If hosts that only contain alphanumerics are allowed ("localhost" for example)
+            # Validate URL
             if not is_safe_valid_url(url):
                 return f"Invalid or unsupported URL - {url}", 400
 
+            # Check for duplicates if dedupe is enabled
             if dedupe and self.datastore.url_exists(url):
                 continue
 
-            new_uuid = self.datastore.add_watch(url=url, extras=extras, tag=tags, tag_uuids=tag_uuids)
-            added.append(new_uuid)
+            urls_to_import.append(url)
 
-        return added
+        # For small imports, process synchronously for immediate feedback
+        if len(urls_to_import) < IMPORT_SWITCH_TO_BACKGROUND_THRESHOLD:
+            added = []
+            for url in urls_to_import:
+                new_uuid = self.datastore.add_watch(url=url, extras=extras, tag=tags, tag_uuids=tag_uuids)
+                added.append(new_uuid)
+            return added, 200
+
+        # For large imports (>= 20), process in background thread
+        else:
+            import threading
+            from loguru import logger
+
+            def import_watches_background():
+                """Background thread to import watches - discarded after completion."""
+                try:
+                    added_count = 0
+                    for url in urls_to_import:
+                        try:
+                            self.datastore.add_watch(url=url, extras=extras, tag=tags, tag_uuids=tag_uuids)
+                            added_count += 1
+                        except Exception as e:
+                            logger.error(f"Error importing URL {url}: {e}")
+
+                    logger.info(f"Background import complete: {added_count} watches created")
+                except Exception as e:
+                    logger.error(f"Error in background import: {e}")
+
+            # Start background thread and return immediately
+            thread = threading.Thread(target=import_watches_background, daemon=True, name="ImportWatches-Background")
+            thread.start()
+
+            return {'status': f'Importing {len(urls_to_import)} URLs in background', 'count': len(urls_to_import)}, 202

changedetectionio/api/Notifications.py

@@ -67,7 +67,7 @@ class Notifications(Resource):
 
         clean_urls = [url.strip() for url in notification_urls if isinstance(url, str)]
         self.datastore.data['settings']['application']['notification_urls'] = clean_urls
-        self.datastore.needs_write = True
+        self.datastore.commit()
 
         return {'notification_urls': clean_urls}, 200
         
@@ -95,7 +95,7 @@ class Notifications(Resource):
             abort(400, message="No matching notification URLs found.")
 
         self.datastore.data['settings']['application']['notification_urls'] = notification_urls
-        self.datastore.needs_write = True
+        self.datastore.commit()
 
         return 'OK', 204
     

changedetectionio/api/Tags.py

@@ -1,5 +1,5 @@
 from changedetectionio import queuedWatchMetaData
-from changedetectionio import worker_handler
+from changedetectionio import worker_pool
 from flask_expects_json import expects_json
 from flask_restful import abort, Resource
 from loguru import logger
@@ -24,8 +24,7 @@ class Tag(Resource):
     @validate_openapi_request('getTag')
     def get(self, uuid):
         """Get data for a single tag/group, toggle notification muting, or recheck all."""
-        from copy import deepcopy
-        tag = deepcopy(self.datastore.data['settings']['application']['tags'].get(uuid))
+        tag = self.datastore.data['settings']['application']['tags'].get(uuid)
         if not tag:
             abort(404, message=f'No tag exists with the UUID of {uuid}')
 
@@ -42,7 +41,7 @@ class Tag(Resource):
             # If less than 20 watches, queue synchronously for immediate feedback
             if len(watches_to_queue) < 20:
                 for watch_uuid in watches_to_queue:
-                    worker_handler.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid}))
+                    worker_pool.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid}))
                 return {'status': f'OK, queued {len(watches_to_queue)} watches for rechecking'}, 200
             else:
                 # 20+ watches - queue in background thread to avoid blocking API response
@@ -50,7 +49,7 @@ class Tag(Resource):
                     """Background thread to queue watches - discarded after completion."""
                     try:
                         for watch_uuid in watches_to_queue:
-                            worker_handler.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid}))
+                            worker_pool.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid}))
                         logger.info(f"Background queueing complete for tag {tag['uuid']}: {len(watches_to_queue)} watches queued")
                     except Exception as e:
                         logger.error(f"Error in background queueing for tag {tag['uuid']}: {e}")
@@ -62,10 +61,12 @@ class Tag(Resource):
                 return {'status': f'OK, queueing {len(watches_to_queue)} watches in background'}, 202
 
         if request.args.get('muted', '') == 'muted':
-            self.datastore.data['settings']['application']['tags'][uuid]['notification_muted'] = True
+            tag['notification_muted'] = True
+            tag.commit()
             return "OK", 200
         elif request.args.get('muted', '') == 'unmuted':
-            self.datastore.data['settings']['application']['tags'][uuid]['notification_muted'] = False
+            tag['notification_muted'] = False
+            tag.commit()
             return "OK", 200
 
         return tag
@@ -79,11 +80,23 @@ class Tag(Resource):
 
         # Delete the tag, and any tag reference
         del self.datastore.data['settings']['application']['tags'][uuid]
-        
+
+        # Delete tag.json file if it exists
+        import os
+        tag_dir = os.path.join(self.datastore.datastore_path, uuid)
+        tag_json = os.path.join(tag_dir, "tag.json")
+        if os.path.exists(tag_json):
+            try:
+                os.unlink(tag_json)
+                logger.info(f"Deleted tag.json for tag {uuid}")
+            except Exception as e:
+                logger.error(f"Failed to delete tag.json for tag {uuid}: {e}")
+
         # Remove tag from all watches
         for watch_uuid, watch in self.datastore.data['watching'].items():
             if watch.get('tags') and uuid in watch['tags']:
                 watch['tags'].remove(uuid)
+                watch.commit()
 
         return 'OK', 204
 
@@ -96,8 +109,18 @@ class Tag(Resource):
         if not tag:
             abort(404, message='No tag exists with the UUID of {}'.format(uuid))
 
+        # Validate notification_urls if provided
+        if 'notification_urls' in request.json:
+            from wtforms import ValidationError
+            from changedetectionio.api.Notifications import validate_notification_urls
+            try:
+                notification_urls = request.json.get('notification_urls', [])
+                validate_notification_urls(notification_urls)
+            except ValidationError as e:
+                return str(e), 400
+
         tag.update(request.json)
-        self.datastore.needs_write_urgent = True
+        tag.commit()
 
         return "OK", 200
 

changedetectionio/api/Watch.py

@@ -6,7 +6,7 @@ from changedetectionio.favicon_utils import get_favicon_mime_type
 
 from . import auth
 from changedetectionio import queuedWatchMetaData, strtobool
-from changedetectionio import worker_handler
+from changedetectionio import worker_pool
 from flask import request, make_response, send_from_directory
 from flask_expects_json import expects_json
 from flask_restful import abort, Resource
@@ -66,47 +66,46 @@ class Watch(Resource):
     @validate_openapi_request('getWatch')
     def get(self, uuid):
         """Get information about a single watch, recheck, pause, or mute."""
-        import time
-        from copy import deepcopy
-        watch = None
-        # Retry up to 20 times if dict is being modified
-        # With sleep(0), this is fast: ~200µs best case, ~20ms worst case under heavy load
-        for attempt in range(20):
-            try:
-                watch = deepcopy(self.datastore.data['watching'].get(uuid))
-                break
-            except RuntimeError:
-                # Dict changed during deepcopy, retry after yielding to scheduler
-                # sleep(0) releases GIL and yields - no fixed delay, just lets other threads run
-                if attempt < 19:  # Don't yield on last attempt
-                    time.sleep(0)  # Yield to scheduler (microseconds, not milliseconds)
-
-        if not watch:
+        # Get watch reference first (for pause/mute operations)
+        watch_obj = self.datastore.data['watching'].get(uuid)
+        if not watch_obj:
             abort(404, message='No watch exists with the UUID of {}'.format(uuid))
 
+        # Create a dict copy for JSON response (with lock for thread safety)
+        # This is much faster than deepcopy and doesn't copy the datastore reference
+        # WARNING: dict() is a SHALLOW copy - nested dicts are shared with original!
+        # Only safe because we only ADD scalar properties (line 97-101), never modify nested dicts
+        # If you need to modify nested dicts, use: from copy import deepcopy; watch = deepcopy(dict(watch_obj))
+        with self.datastore.lock:
+            watch = dict(watch_obj)
+
         if request.args.get('recheck'):
-            worker_handler.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+            worker_pool.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
             return "OK", 200
         if request.args.get('paused', '') == 'paused':
-            self.datastore.data['watching'].get(uuid).pause()
+            watch_obj.pause()
+            watch_obj.commit()
             return "OK", 200
         elif request.args.get('paused', '') == 'unpaused':
-            self.datastore.data['watching'].get(uuid).unpause()
+            watch_obj.unpause()
+            watch_obj.commit()
             return "OK", 200
         if request.args.get('muted', '') == 'muted':
-            self.datastore.data['watching'].get(uuid).mute()
+            watch_obj.mute()
+            watch_obj.commit()
             return "OK", 200
         elif request.args.get('muted', '') == 'unmuted':
-            self.datastore.data['watching'].get(uuid).unmute()
+            watch_obj.unmute()
+            watch_obj.commit()
             return "OK", 200
 
         # Return without history, get that via another API call
         # Properties are not returned as a JSON, so add the required props manually
-        watch['history_n'] = watch.history_n
+        watch['history_n'] = watch_obj.history_n
         # attr .last_changed will check for the last written text snapshot on change
-        watch['last_changed'] = watch.last_changed
-        watch['viewed'] = watch.viewed
-        watch['link'] = watch.link,
+        watch['last_changed'] = watch_obj.last_changed
+        watch['viewed'] = watch_obj.viewed
+        watch['link'] = watch_obj.link,
 
         return watch
 
@@ -140,6 +139,16 @@ class Watch(Resource):
         if validation_error:
             return validation_error, 400
 
+        # Validate notification_urls if provided
+        if 'notification_urls' in request.json:
+            from wtforms import ValidationError
+            from changedetectionio.api.Notifications import validate_notification_urls
+            try:
+                notification_urls = request.json.get('notification_urls', [])
+                validate_notification_urls(notification_urls)
+            except ValidationError as e:
+                return str(e), 400
+
         # XSS etc protection - validate URL if it's being updated
         if 'url' in request.json:
             new_url = request.json.get('url')
@@ -159,58 +168,19 @@ class Watch(Resource):
 
         # Handle processor-config-* fields separately (save to JSON, not datastore)
         from changedetectionio import processors
-        processor_config_data = {}
-        regular_data = {}
 
-        for key, value in request.json.items():
-            if key.startswith('processor_config_'):
-                config_key = key.replace('processor_config_', '')
-                if value:  # Only save non-empty values
-                    processor_config_data[config_key] = value
-            else:
-                regular_data[key] = value
+        # Make a mutable copy of request.json for modification
+        json_data = dict(request.json)
+
+        # Extract and remove processor config fields from json_data
+        processor_config_data = processors.extract_processor_config_from_form_data(json_data)
 
         # Update watch with regular (non-processor-config) fields
-        watch.update(regular_data)
+        watch.update(json_data)
+        watch.commit()
 
-        # Save processor config to JSON file if any config data exists
-        if processor_config_data:
-            try:
-                processor_name = request.json.get('processor', watch.get('processor'))
-                if processor_name:
-                    # Create a processor instance to access config methods
-                    from changedetectionio.processors import difference_detection_processor
-                    processor_instance = difference_detection_processor(self.datastore, uuid)
-                    # Use processor name as filename so each processor keeps its own config
-                    config_filename = f'{processor_name}.json'
-                    processor_instance.update_extra_watch_config(config_filename, processor_config_data)
-                    logger.debug(f"API: Saved processor config to {config_filename}: {processor_config_data}")
-
-                    # Call optional edit_hook if processor has one
-                    try:
-                        import importlib
-                        edit_hook_module_name = f'changedetectionio.processors.{processor_name}.edit_hook'
-
-                        try:
-                            edit_hook = importlib.import_module(edit_hook_module_name)
-                            logger.debug(f"API: Found edit_hook module for {processor_name}")
-
-                            if hasattr(edit_hook, 'on_config_save'):
-                                logger.info(f"API: Calling edit_hook.on_config_save for {processor_name}")
-                                # Call hook and get updated config
-                                updated_config = edit_hook.on_config_save(watch, processor_config_data, self.datastore)
-                                # Save updated config back to file
-                                processor_instance.update_extra_watch_config(config_filename, updated_config)
-                                logger.info(f"API: Edit hook updated config: {updated_config}")
-                            else:
-                                logger.debug(f"API: Edit hook module found but no on_config_save function")
-                        except ModuleNotFoundError:
-                            logger.debug(f"API: No edit_hook module for processor {processor_name} (this is normal)")
-                    except Exception as hook_error:
-                        logger.error(f"API: Edit hook error (non-fatal): {hook_error}", exc_info=True)
-
-            except Exception as e:
-                logger.error(f"API: Failed to save processor config: {e}")
+        # Save processor config to JSON file
+        processors.save_processor_config(self.datastore, uuid, processor_config_data)
 
         return "OK", 200
 
@@ -404,10 +374,10 @@ class WatchFavicon(Resource):
         favicon_filename = watch.get_favicon_filename()
         if favicon_filename:
             # Use cached MIME type detection
-            filepath = os.path.join(watch.watch_data_dir, favicon_filename)
+            filepath = os.path.join(watch.data_dir, favicon_filename)
             mime = get_favicon_mime_type(filepath)
 
-            response = make_response(send_from_directory(watch.watch_data_dir, favicon_filename))
+            response = make_response(send_from_directory(watch.data_dir, favicon_filename))
             response.headers['Content-type'] = mime
             response.headers['Cache-Control'] = 'max-age=300, must-revalidate'  # Cache for 5 minutes, then revalidate
             return response
@@ -444,8 +414,24 @@ class CreateWatch(Resource):
         if validation_error:
             return validation_error, 400
 
+        # Validate notification_urls if provided
+        if 'notification_urls' in json_data:
+            from wtforms import ValidationError
+            from changedetectionio.api.Notifications import validate_notification_urls
+            try:
+                notification_urls = json_data.get('notification_urls', [])
+                validate_notification_urls(notification_urls)
+            except ValidationError as e:
+                return str(e), 400
+
+        # Handle processor-config-* fields separately (save to JSON, not watch)
+        from changedetectionio import processors
+
         extras = copy.deepcopy(json_data)
 
+        # Extract and remove processor config fields from extras
+        processor_config_data = processors.extract_processor_config_from_form_data(extras)
+
         # Because we renamed 'tag' to 'tags' but don't want to change the API (can do this in v2 of the API)
         tags = None
         if extras.get('tag'):
@@ -455,11 +441,25 @@ class CreateWatch(Resource):
         del extras['url']
 
         new_uuid = self.datastore.add_watch(url=url, extras=extras, tag=tags)
+
+        # Save processor config to separate JSON file
+        if new_uuid and processor_config_data:
+            processors.save_processor_config(self.datastore, new_uuid, processor_config_data)
         if new_uuid:
 # Dont queue because the scheduler will check that it hasnt been checked before anyway
-#            worker_handler.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': new_uuid}))
+#            worker_pool.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': new_uuid}))
             return {'uuid': new_uuid}, 201
         else:
+            # Check if it was a limit issue
+            page_watch_limit = os.getenv('PAGE_WATCH_LIMIT')
+            if page_watch_limit:
+                try:
+                    page_watch_limit = int(page_watch_limit)
+                    current_watch_count = len(self.datastore.data['watching'])
+                    if current_watch_count >= page_watch_limit:
+                        return f"Watch limit reached ({current_watch_count}/{page_watch_limit} watches). Cannot add more watches.", 429
+                except ValueError:
+                    pass
             return "Invalid or unsupported URL", 400
 
     @auth.check_token
@@ -481,6 +481,7 @@ class CreateWatch(Resource):
                 'last_error': watch['last_error'],
                 'link': watch.link,
                 'page_title': watch['page_title'],
+                'tags': [*tags],  # Unpack dict keys to list (can't use list() since variable named 'list')
                 'title': watch['title'],
                 'url': watch['url'],
                 'viewed': watch.viewed
@@ -494,7 +495,7 @@ class CreateWatch(Resource):
             if len(watches_to_queue) < 20:
                 # Get already queued/running UUIDs once (efficient)
                 queued_uuids = set(self.update_q.get_queued_uuids())
-                running_uuids = set(worker_handler.get_running_uuids())
+                running_uuids = set(worker_pool.get_running_uuids())
 
                 # Filter out watches that are already queued or running
                 watches_to_queue_filtered = [
@@ -504,7 +505,7 @@ class CreateWatch(Resource):
 
                 # Queue only the filtered watches
                 for uuid in watches_to_queue_filtered:
-                    worker_handler.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+                    worker_pool.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
 
                 # Provide feedback about skipped watches
                 skipped_count = len(watches_to_queue) - len(watches_to_queue_filtered)
@@ -516,7 +517,7 @@ class CreateWatch(Resource):
                 # 20+ watches - queue in background thread to avoid blocking API response
                 # Capture queued/running state before background thread
                 queued_uuids = set(self.update_q.get_queued_uuids())
-                running_uuids = set(worker_handler.get_running_uuids())
+                running_uuids = set(worker_pool.get_running_uuids())
 
                 def queue_all_watches_background():
                     """Background thread to queue all watches - discarded after completion."""
@@ -526,7 +527,7 @@ class CreateWatch(Resource):
                         for uuid in watches_to_queue:
                             # Check if already queued or running (state captured at start)
                             if uuid not in queued_uuids and uuid not in running_uuids:
-                                worker_handler.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+                                worker_pool.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
                                 queued_count += 1
                             else:
                                 skipped_count += 1

changedetectionio/api/__init__.py

@@ -12,9 +12,17 @@ schema = api_schema.build_watch_json_schema(watch_base_config)
 schema_create_watch = copy.deepcopy(schema)
 schema_create_watch['required'] = ['url']
 del schema_create_watch['properties']['last_viewed']
+# Allow processor_config_* fields (handled separately in endpoint)
+schema_create_watch['patternProperties'] = {
+    '^processor_config_': {'type': ['string', 'number', 'boolean', 'object', 'array', 'null']}
+}
 
 schema_update_watch = copy.deepcopy(schema)
 schema_update_watch['additionalProperties'] = False
+# Allow processor_config_* fields (handled separately in endpoint)
+schema_update_watch['patternProperties'] = {
+    '^processor_config_': {'type': ['string', 'number', 'boolean', 'object', 'array', 'null']}
+}
 
 # Tag schema is also based on watch_base since Tag inherits from it
 schema_tag = copy.deepcopy(schema)

changedetectionio/blueprint/backups/__init__.py

@@ -27,15 +27,27 @@ def create_backup(datastore_path, watches: dict):
                          compression=zipfile.ZIP_DEFLATED,
                          compresslevel=8) as zipObj:
 
-        # Add the index
-        zipObj.write(os.path.join(datastore_path, "url-watches.json"), arcname="url-watches.json")
+        # Add the settings file (supports both formats)
+        # New format: changedetection.json
+        changedetection_json = os.path.join(datastore_path, "changedetection.json")
+        if os.path.isfile(changedetection_json):
+            zipObj.write(changedetection_json, arcname="changedetection.json")
+            logger.debug("Added changedetection.json to backup")
 
-        # Add the flask app secret
-        zipObj.write(os.path.join(datastore_path, "secret.txt"), arcname="secret.txt")
+        # Legacy format: url-watches.json (for backward compatibility)
+        url_watches_json = os.path.join(datastore_path, "url-watches.json")
+        if os.path.isfile(url_watches_json):
+            zipObj.write(url_watches_json, arcname="url-watches.json")
+            logger.debug("Added url-watches.json to backup")
+
+        # Add the flask app secret (if it exists)
+        secret_file = os.path.join(datastore_path, "secret.txt")
+        if os.path.isfile(secret_file):
+            zipObj.write(secret_file, arcname="secret.txt")
 
         # Add any data in the watch data directory.
         for uuid, w in watches.items():
-            for f in Path(w.watch_data_dir).glob('*'):
+            for f in Path(w.data_dir).glob('*'):
                 zipObj.write(f,
                              # Use the full path to access the file, but make the file 'relative' in the Zip.
                              arcname=os.path.join(f.parts[-2], f.parts[-1]),
@@ -90,8 +102,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
             flash(gettext("Maximum number of backups reached, please remove some"), "error")
             return redirect(url_for('backups.index'))
 
-        # Be sure we're written fresh
-        datastore.sync_to_json()
+        # With immediate persistence, all data is already saved
         zip_thread = threading.Thread(
             target=create_backup,
             args=(datastore.datastore_path, datastore.data.get("watching")),

changedetectionio/blueprint/backups/templates/overview.html

@@ -3,10 +3,10 @@
     {% from '_helpers.html' import render_simple_field, render_field %}
     <div class="edit-form">
         <div class="box-wrap inner">
-            <h4>{{ _('Backups') }}</h4>
+            <h2>{{ _('Backups') }}</h2>
             {% if backup_running %}
                 <p>
-                    <strong>{{ _('A backup is running!') }}</strong>
+                    <span class="spinner"></span>&nbsp;<strong>{{ _('A backup is running!') }}</strong>
                 </p>
             {% endif %}
             <p>

changedetectionio/blueprint/browser_steps/__init__.py

@@ -285,8 +285,8 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         watch = datastore.data['watching'].get(uuid)
         filename = f"step_before-{step_n}.jpeg" if request.args.get('type', '') == 'before' else f"step_{step_n}.jpeg"
 
-        if step_n and watch and os.path.isfile(os.path.join(watch.watch_data_dir, filename)):
-            response = make_response(send_from_directory(directory=watch.watch_data_dir, path=filename))
+        if step_n and watch and os.path.isfile(os.path.join(watch.data_dir, filename)):
+            response = make_response(send_from_directory(directory=watch.data_dir, path=filename))
             response.headers['Content-type'] = 'image/jpeg'
             response.headers['Cache-Control'] = 'no-cache, no-store, must-revalidate'
             response.headers['Pragma'] = 'no-cache'

changedetectionio/blueprint/imports/__init__.py

@@ -14,7 +14,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
         from changedetectionio import forms
 #
         if request.method == 'POST':
-#            from changedetectionio import worker_handler
+#            from changedetectionio import worker_pool
 
             from changedetectionio.blueprint.imports.importer import (
                 import_url_list,
@@ -26,12 +26,13 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
             # URL List import
             if request.values.get('urls') and len(request.values.get('urls').strip()):
                 # Import and push into the queue for immediate update check
+                from changedetectionio import processors
                 importer_handler = import_url_list()
-                importer_handler.run(data=request.values.get('urls'), flash=flash, datastore=datastore, processor=request.values.get('processor', 'text_json_diff'))
+                importer_handler.run(data=request.values.get('urls'), flash=flash, datastore=datastore, processor=request.values.get('processor', processors.get_default_processor()))
                 logger.debug(f"Imported {len(importer_handler.new_uuids)} new UUIDs")
                 # Dont' add to queue because scheduler can see that they haven't been checked and will add them to the queue
 #                for uuid in importer_handler.new_uuids:
-#                    worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+#                    worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
 
                 if len(importer_handler.remaining_data) == 0:
                     return redirect(url_for('watchlist.index'))
@@ -45,7 +46,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
                 d_importer.run(data=request.values.get('distill-io'), flash=flash, datastore=datastore)
                 # Dont' add to queue because scheduler can see that they haven't been checked and will add them to the queue
 #                for uuid in importer_handler.new_uuids:
-#                    worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+#                    worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
 
 
             # XLSX importer
@@ -70,7 +71,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
 
                 # Dont' add to queue because scheduler can see that they haven't been checked and will add them to the queue
 #                for uuid in importer_handler.new_uuids:
-#                    worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+#                    worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
 
 
         # Could be some remaining, or we could be on GET

changedetectionio/blueprint/imports/importer.py

@@ -62,7 +62,7 @@ class import_url_list(Importer):
                 extras = None
                 if processor:
                     extras = {'processor': processor}
-                new_uuid = datastore.add_watch(url=url.strip(), tag=tags, write_to_disk_now=False, extras=extras)
+                new_uuid = datastore.add_watch(url=url.strip(), tag=tags, save_immediately=False, extras=extras)
 
                 if new_uuid:
                     # Straight into the queue.
@@ -129,7 +129,7 @@ class import_distill_io_json(Importer):
                 new_uuid = datastore.add_watch(url=d['uri'].strip(),
                                                tag=",".join(d.get('tags', [])),
                                                extras=extras,
-                                               write_to_disk_now=False)
+                                               save_immediately=False)
 
                 if new_uuid:
                     # Straight into the queue.
@@ -204,7 +204,7 @@ class import_xlsx_wachete(Importer):
                     new_uuid = datastore.add_watch(url=data['url'].strip(),
                                                    extras=extras,
                                                    tag=data.get('folder'),
-                                                   write_to_disk_now=False)
+                                                   save_immediately=False)
                     if new_uuid:
                         # Straight into the queue.
                         self.new_uuids.append(new_uuid)
@@ -287,7 +287,7 @@ class import_xlsx_custom(Importer):
                     new_uuid = datastore.add_watch(url=url,
                                                    extras=extras,
                                                    tag=tags,
-                                                   write_to_disk_now=False)
+                                                   save_immediately=False)
                     if new_uuid:
                         # Straight into the queue.
                         self.new_uuids.append(new_uuid)

changedetectionio/blueprint/price_data_follower/__init__.py

@@ -4,7 +4,7 @@ from flask import Blueprint, flash, redirect, url_for
 from flask_login import login_required
 from changedetectionio.store import ChangeDetectionStore
 from changedetectionio import queuedWatchMetaData
-from changedetectionio import worker_handler
+from changedetectionio import worker_pool
 from queue import PriorityQueue
 
 PRICE_DATA_TRACK_ACCEPT = 'accepted'
@@ -20,13 +20,15 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q: PriorityQueue
         datastore.data['watching'][uuid]['track_ldjson_price_data'] = PRICE_DATA_TRACK_ACCEPT
         datastore.data['watching'][uuid]['processor'] = 'restock_diff'
         datastore.data['watching'][uuid].clear_watch()
-        worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+        datastore.data['watching'][uuid].commit()
+        worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
         return redirect(url_for("watchlist.index"))
 
     @login_required
     @price_data_follower_blueprint.route("/<string:uuid>/reject", methods=['GET'])
     def reject(uuid):
         datastore.data['watching'][uuid]['track_ldjson_price_data'] = PRICE_DATA_TRACK_REJECT
+        datastore.data['watching'][uuid].commit()
         return redirect(url_for("watchlist.index"))
 
 

changedetectionio/blueprint/rss/single_watch.py

@@ -37,6 +37,8 @@ def construct_single_watch_routes(rss_blueprint, datastore):
 
         rss_content_format = datastore.data['settings']['application'].get('rss_content_format')
 
+        if uuid == 'first':
+            uuid = list(datastore.data['watching'].keys()).pop()
         # Get the watch by UUID
         watch = datastore.data['watching'].get(uuid)
         if not watch:

changedetectionio/blueprint/settings/__init__.py

@@ -1,8 +1,9 @@
 import os
 from copy import deepcopy
-from datetime import datetime
+from datetime import datetime, timedelta
 from zoneinfo import ZoneInfo, available_timezones
 import secrets
+import time
 import flask_login
 from flask import Blueprint, render_template, request, redirect, url_for, flash
 from flask_babel import gettext
@@ -74,16 +75,17 @@ def construct_blueprint(datastore: ChangeDetectionStore):
                     del (app_update['password'])
 
                 datastore.data['settings']['application'].update(app_update)
-                
+
                 # Handle dynamic worker count adjustment
                 old_worker_count = datastore.data['settings']['requests'].get('workers', 1)
                 new_worker_count = form.data['requests'].get('workers', 1)
 
                 datastore.data['settings']['requests'].update(form.data['requests'])
+                datastore.commit()
 
                 # Adjust worker count if it changed
                 if new_worker_count != old_worker_count:
-                    from changedetectionio import worker_handler
+                    from changedetectionio import worker_pool
                     from changedetectionio.flask_app import update_q, notification_q, app, datastore as ds
 
                     # Check CPU core availability and warn if worker count is high
@@ -92,7 +94,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
                         flash(gettext("Warning: Worker count ({}) is close to or exceeds available CPU cores ({})").format(
                             new_worker_count, cpu_count), 'warning')
 
-                    result = worker_handler.adjust_async_worker_count(
+                    result = worker_pool.adjust_async_worker_count(
                         new_count=new_worker_count,
                         update_q=update_q,
                         notification_q=notification_q,
@@ -109,13 +111,11 @@ def construct_blueprint(datastore: ChangeDetectionStore):
 
                 if not os.getenv("SALTED_PASS", False) and len(form.application.form.password.encrypted_password):
                     datastore.data['settings']['application']['password'] = form.application.form.password.encrypted_password
-                    datastore.needs_write_urgent = True
+                    datastore.commit()
                     flash(gettext("Password protection enabled."), 'notice')
                     flask_login.logout_user()
                     return redirect(url_for('watchlist.index'))
 
-                datastore.needs_write_urgent = True
-
                 # Also save plugin settings from the same form submission
                 plugin_tabs_list = get_plugin_settings_tabs()
                 for tab in plugin_tabs_list:
@@ -143,6 +143,9 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         active_plugins = get_active_plugins()
         python_version = f"{sys.version_info.major}.{sys.version_info.minor}.{sys.version_info.micro}"
 
+        # Calculate uptime in seconds
+        uptime_seconds = time.time() - datastore.start_time
+
         # Get plugin settings tabs and instantiate forms
         plugin_tabs = get_plugin_settings_tabs()
         plugin_forms = {}
@@ -161,6 +164,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
                                 active_plugins=active_plugins,
                                 api_key=datastore.data['settings']['application'].get('api_access_token'),
                                 python_version=python_version,
+                                uptime_seconds=uptime_seconds,
                                 available_timezones=sorted(available_timezones()),
                                 emailprefix=os.getenv('NOTIFICATION_MAIL_BUTTON_PREFIX', False),
                                 extra_notification_token_placeholder_info=datastore.get_unique_notification_token_placeholders_available(),
@@ -181,7 +185,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     def settings_reset_api_key():
         secret = secrets.token_hex(16)
         datastore.data['settings']['application']['api_access_token'] = secret
-        datastore.needs_write_urgent = True
+        datastore.commit()
         flash(gettext("API Key was regenerated."))
         return redirect(url_for('settings.settings_page')+'#api')
         
@@ -198,7 +202,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     def toggle_all_paused():
         current_state = datastore.data['settings']['application'].get('all_paused', False)
         datastore.data['settings']['application']['all_paused'] = not current_state
-        datastore.needs_write_urgent = True
+        datastore.commit()
 
         if datastore.data['settings']['application']['all_paused']:
             flash(gettext("Automatic scheduling paused - checks will not be queued."), 'notice')
@@ -212,7 +216,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     def toggle_all_muted():
         current_state = datastore.data['settings']['application'].get('all_muted', False)
         datastore.data['settings']['application']['all_muted'] = not current_state
-        datastore.needs_write_urgent = True
+        datastore.commit()
 
         if datastore.data['settings']['application']['all_muted']:
             flash(gettext("All notifications muted."), 'notice')

changedetectionio/blueprint/settings/templates/settings.html

@@ -25,7 +25,7 @@
             <li class="tab"><a href="#ui-options">{{ _('UI Options') }}</a></li>
             <li class="tab"><a href="#api">{{ _('API') }}</a></li>
             <li class="tab"><a href="#rss">{{ _('RSS') }}</a></li>
-            <li class="tab"><a href="{{ url_for('backups.index') }}" class="pure-menu-link">{{ _('Backups') }}</a></li>
+            <li class="tab"><a href="{{ url_for('backups.index') }}">{{ _('Backups') }}</a></li>
             <li class="tab"><a href="#timedate">{{ _('Time & Date') }}</a></li>
             <li class="tab"><a href="#proxies">{{ _('CAPTCHA & Proxies') }}</a></li>
             {% if plugin_tabs %}
@@ -59,6 +59,14 @@
                         {{ _('Set to') }} <strong>0</strong> {{ _('to disable') }}
                         </span>
                     </div>
+                    <div class="pure-control-group">
+                        {{ render_field(form.application.form.history_snapshot_max_length, class="history_snapshot_max_length") }}
+                        <span class="pure-form-message-inline">{{ _('Limit collection of history snapshots for each watch to this number of history items.') }}
+                            <br>
+                        {{ _('Set to empty to disable / no limit') }}
+                        </span>
+                    </div>
+
                     <div class="pure-control-group">
                         {% if not hide_remove_pass %}
                             {% if current_user.is_authenticated %}
@@ -80,6 +88,16 @@
                         {{ render_checkbox_field(form.application.form.empty_pages_are_a_change) }}
                         <span class="pure-form-message-inline">{{ _('When a request returns no content, or the HTML does not contain any text, is this considered a change?') }}</span>
                     </div>
+                {% if form.requests.proxy %}
+                    <div>
+                    <br>
+                        <div class="inline-radio">
+                            {{ render_field(form.requests.form.proxy, class="fetch-backend-proxy") }}
+                            <span class="pure-form-message-inline">{{ _('Choose a default proxy for all watches') }}</span>
+                        </div>
+                    </div>
+                {% endif %}
+
                 </fieldset>
             </div>
 
@@ -340,15 +358,6 @@ nav
                 {{ render_fieldlist_with_inline_errors(form.requests.form.extra_proxies) }}
                 <span class="pure-form-message-inline">{{ _('"Name" will be used for selecting the proxy in the Watch Edit settings') }}</span><br>
                 <span class="pure-form-message-inline">{{ _('SOCKS5 proxies with authentication are only supported with \'plain requests\' fetcher, for other fetchers you should whitelist the IP access instead') }}</span>
-                {% if form.requests.proxy %}
-                <div>
-                <br>
-                    <div class="inline-radio">
-                        {{ render_field(form.requests.form.proxy, class="fetch-backend-proxy") }}
-                        <span class="pure-form-message-inline">{{ _('Choose a default proxy for all watches') }}</span>
-                    </div>
-                </div>
-                {% endif %}
                 </div>
                 <div class="pure-control-group" id="extra-browsers-setting">
                     <p>
@@ -385,6 +394,7 @@ nav
                 {% endfor %}
             {% endif %}
             <div class="tab-pane-inner" id="info">
+                <p><strong>{{ _('Uptime:') }}</strong> {{ uptime_seconds|format_duration }}</p>
                 <p><strong>{{ _('Python version:') }}</strong> {{ python_version }}</p>
                 <p><strong>{{ _('Plugins active:') }}</strong></p>
                 {% if active_plugins %}

changedetectionio/blueprint/tags/__init__.py

@@ -57,8 +57,10 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     @tags_blueprint.route("/mute/<string:uuid>", methods=['GET'])
     @login_optionally_required
     def mute(uuid):
-        if datastore.data['settings']['application']['tags'].get(uuid):
-            datastore.data['settings']['application']['tags'][uuid]['notification_muted'] = not datastore.data['settings']['application']['tags'][uuid]['notification_muted']
+        tag = datastore.data['settings']['application']['tags'].get(uuid)
+        if tag:
+            tag['notification_muted'] = not tag['notification_muted']
+            tag.commit()
         return redirect(url_for('tags.tags_overview_page'))
 
     @tags_blueprint.route("/delete/<string:uuid>", methods=['GET'])
@@ -68,6 +70,17 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         if datastore.data['settings']['application']['tags'].get(uuid):
             del datastore.data['settings']['application']['tags'][uuid]
 
+        # Delete tag.json file if it exists
+        import os
+        tag_dir = os.path.join(datastore.datastore_path, uuid)
+        tag_json = os.path.join(tag_dir, "tag.json")
+        if os.path.exists(tag_json):
+            try:
+                os.unlink(tag_json)
+                logger.info(f"Deleted tag.json for tag {uuid}")
+            except Exception as e:
+                logger.error(f"Failed to delete tag.json for tag {uuid}: {e}")
+
         # Remove tag from all watches in background thread to avoid blocking
         def remove_tag_background(tag_uuid):
             """Background thread to remove tag from watches - discarded after completion."""
@@ -76,6 +89,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
                 for watch_uuid, watch in datastore.data['watching'].items():
                     if watch.get('tags') and tag_uuid in watch['tags']:
                         watch['tags'].remove(tag_uuid)
+                        watch.commit()
                         removed_count += 1
                 logger.info(f"Background: Tag {tag_uuid} removed from {removed_count} watches")
             except Exception as e:
@@ -98,6 +112,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
                 for watch_uuid, watch in datastore.data['watching'].items():
                     if watch.get('tags') and tag_uuid in watch['tags']:
                         watch['tags'].remove(tag_uuid)
+                        watch.commit()
                         unlinked_count += 1
                 logger.info(f"Background: Tag {tag_uuid} unlinked from {unlinked_count} watches")
             except Exception as e:
@@ -112,6 +127,17 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     @tags_blueprint.route("/delete_all", methods=['GET'])
     @login_optionally_required
     def delete_all():
+        # Delete all tag.json files
+        import os
+        for tag_uuid in list(datastore.data['settings']['application']['tags'].keys()):
+            tag_dir = os.path.join(datastore.datastore_path, tag_uuid)
+            tag_json = os.path.join(tag_dir, "tag.json")
+            if os.path.exists(tag_json):
+                try:
+                    os.unlink(tag_json)
+                except Exception as e:
+                    logger.error(f"Failed to delete tag.json for tag {tag_uuid}: {e}")
+
         # Clear all tags from settings immediately
         datastore.data['settings']['application']['tags'] = {}
 
@@ -122,6 +148,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
             try:
                 for watch_uuid, watch in datastore.data['watching'].items():
                     watch['tags'] = []
+                    watch.commit()
                     cleared_count += 1
                 logger.info(f"Background: Cleared tags from {cleared_count} watches")
             except Exception as e:
@@ -202,10 +229,10 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         if uuid == 'first':
             uuid = list(datastore.data['settings']['application']['tags'].keys()).pop()
 
-        default = datastore.data['settings']['application']['tags'].get(uuid)
+        tag = datastore.data['settings']['application']['tags'].get(uuid)
 
         form = group_restock_settings_form(formdata=request.form if request.method == 'POST' else None,
-                               data=default,
+                               data=tag,
                                extra_notification_tokens=datastore.get_unique_notification_tokens_available()
                                )
         # @todo subclass form so validation works
@@ -214,9 +241,9 @@ def construct_blueprint(datastore: ChangeDetectionStore):
 #                flash(','.join(l), 'error')
 #           return redirect(url_for('tags.form_tag_edit_submit', uuid=uuid))
 
-        datastore.data['settings']['application']['tags'][uuid].update(form.data)
-        datastore.data['settings']['application']['tags'][uuid]['processor'] = 'restock_diff'
-        datastore.needs_write_urgent = True
+        tag.update(form.data)
+        tag['processor'] = 'restock_diff'
+        tag.commit()
         flash(gettext("Updated"))
 
         return redirect(url_for('tags.tags_overview_page'))

changedetectionio/blueprint/ui/__init__.py

@@ -1,6 +1,6 @@
 import time
 import threading
-from flask import Blueprint, request, redirect, url_for, flash, render_template, session
+from flask import Blueprint, request, redirect, url_for, flash, render_template, session, current_app
 from flask_babel import gettext
 from loguru import logger
 
@@ -10,7 +10,7 @@ from changedetectionio.blueprint.ui.notification import construct_blueprint as c
 from changedetectionio.blueprint.ui.views import construct_blueprint as construct_views_blueprint
 from changedetectionio.blueprint.ui import diff, preview
 
-def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWatchMetaData, watch_check_update, extra_data=None, emit_flash=True):
+def _handle_operations(op, uuids, datastore, worker_pool, update_q, queuedWatchMetaData, watch_check_update, extra_data=None, emit_flash=True):
     from flask import request, flash
 
     if op == 'delete':
@@ -24,6 +24,7 @@ def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWat
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid]['paused'] = True
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches paused").format(len(uuids)))
 
@@ -31,6 +32,7 @@ def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWat
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid.strip()]['paused'] = False
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches unpaused").format(len(uuids)))
 
@@ -45,6 +47,7 @@ def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWat
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid]['notification_muted'] = True
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches muted").format(len(uuids)))
 
@@ -52,6 +55,7 @@ def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWat
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid]['notification_muted'] = False
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches un-muted").format(len(uuids)))
 
@@ -59,7 +63,7 @@ def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWat
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 # Recheck and require a full reprocessing
-                worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+                worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
         if emit_flash:
             flash(gettext("{} watches queued for rechecking").format(len(uuids)))
 
@@ -67,6 +71,7 @@ def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWat
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid]["last_error"] = False
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches errors cleared").format(len(uuids)))
 
@@ -87,6 +92,7 @@ def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWat
                 datastore.data['watching'][uuid]['notification_body'] = None
                 datastore.data['watching'][uuid]['notification_urls'] = []
                 datastore.data['watching'][uuid]['notification_format'] = USE_SYSTEM_DEFAULT_NOTIFICATION_FORMAT_FOR_WATCH
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches set to use default notification settings").format(len(uuids)))
 
@@ -102,6 +108,7 @@ def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWat
                             datastore.data['watching'][uuid]['tags'] = []
 
                         datastore.data['watching'][uuid]['tags'].append(tag_uuid)
+                        datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches were tagged").format(len(uuids)))
 
@@ -109,7 +116,7 @@ def _handle_operations(op, uuids, datastore, worker_handler, update_q, queuedWat
         for uuid in uuids:
             watch_check_update.send(watch_uuid=uuid)
 
-def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handler, queuedWatchMetaData, watch_check_update):
+def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_pool, queuedWatchMetaData, watch_check_update):
     ui_blueprint = Blueprint('ui', __name__, template_folder="templates")
     
     # Register the edit blueprint
@@ -217,14 +224,14 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
     @login_optionally_required
     def form_delete():
         uuid = request.args.get('uuid')
+        # More for testing, possible to return the first/only
+        if uuid == 'first':
+            uuid = list(datastore.data['watching'].keys()).pop()
 
         if uuid != 'all' and not uuid in datastore.data['watching'].keys():
             flash(gettext('The watch by UUID {} does not exist.').format(uuid), 'error')
             return redirect(url_for('watchlist.index'))
 
-        # More for testing, possible to return the first/only
-        if uuid == 'first':
-            uuid = list(datastore.data['watching'].keys()).pop()
         datastore.delete(uuid)
         flash(gettext('Deleted.'))
 
@@ -234,14 +241,14 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
     @login_optionally_required
     def form_clone():
         uuid = request.args.get('uuid')
-        # More for testing, possible to return the first/only
+
         if uuid == 'first':
             uuid = list(datastore.data['watching'].keys()).pop()
 
         new_uuid = datastore.clone(uuid)
 
         if not datastore.data['watching'].get(uuid).get('paused'):
-            worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=5, item={'uuid': new_uuid}))
+            worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=5, item={'uuid': new_uuid}))
 
         flash(gettext('Cloned, you are editing the new watch.'))
 
@@ -257,10 +264,10 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
 
         if uuid:
             # Single watch - check if already queued or running
-            if worker_handler.is_watch_running(uuid) or uuid in update_q.get_queued_uuids():
+            if worker_pool.is_watch_running(uuid) or uuid in update_q.get_queued_uuids():
                 flash(gettext("Watch is already queued or being checked."))
             else:
-                worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+                worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
                 flash(gettext("Queued 1 watch for rechecking."))
         else:
             # Multiple watches - first count how many need to be queued
@@ -279,7 +286,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
             if len(watches_to_queue) < 20:
                 # Get already queued/running UUIDs once (efficient)
                 queued_uuids = set(update_q.get_queued_uuids())
-                running_uuids = set(worker_handler.get_running_uuids())
+                running_uuids = set(worker_pool.get_running_uuids())
 
                 # Filter out watches that are already queued or running
                 watches_to_queue_filtered = []
@@ -289,7 +296,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
 
                 # Queue only the filtered watches
                 for watch_uuid in watches_to_queue_filtered:
-                    worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid}))
+                    worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid}))
 
                 # Provide feedback about skipped watches
                 skipped_count = len(watches_to_queue) - len(watches_to_queue_filtered)
@@ -305,7 +312,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
                 # 20+ watches - queue in background thread to avoid blocking HTTP response
                 # Capture queued/running state before background thread
                 queued_uuids = set(update_q.get_queued_uuids())
-                running_uuids = set(worker_handler.get_running_uuids())
+                running_uuids = set(worker_pool.get_running_uuids())
 
                 def queue_watches_background():
                     """Background thread to queue watches - discarded after completion."""
@@ -315,7 +322,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
                         for watch_uuid in watches_to_queue:
                             # Check if already queued or running (state captured at start)
                             if watch_uuid not in queued_uuids and watch_uuid not in running_uuids:
-                                worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid}))
+                                worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': watch_uuid}))
                                 queued_count += 1
                             else:
                                 skipped_count += 1
@@ -344,7 +351,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
             extra_data=extra_data,
             queuedWatchMetaData=queuedWatchMetaData,
             uuids=uuids,
-            worker_handler=worker_handler,
+            worker_pool=worker_pool,
             update_q=update_q,
             watch_check_update=watch_check_update,
             op=op,
@@ -362,9 +369,6 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
         import json
         from copy import deepcopy
 
-        # more for testing
-        if uuid == 'first':
-            uuid = list(datastore.data['watching'].keys()).pop()
 
         # copy it to memory as trim off what we dont need (history)
         watch = deepcopy(datastore.data['watching'].get(uuid))
@@ -404,4 +408,25 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, worker_handle
 
         return redirect(url_for('watchlist.index'))
 
+    @ui_blueprint.route("/language/auto-detect", methods=['GET'])
+    def delete_locale_language_session_var_if_it_exists():
+        """Clear the session locale preference to auto-detect from browser Accept-Language header"""
+        if 'locale' in session:
+            session.pop('locale', None)
+            # Refresh Flask-Babel to clear cached locale
+            from flask_babel import refresh
+            refresh()
+            flash(gettext("Language set to auto-detect from browser"))
+
+        # Check if there's a redirect parameter to return to the same page
+        redirect_url = request.args.get('redirect')
+
+        # If redirect is provided and safe, use it
+        from changedetectionio.is_safe_url import is_safe_url
+        if redirect_url and is_safe_url(redirect_url, current_app):
+            return redirect(redirect_url)
+
+        # Otherwise redirect to watchlist
+        return redirect(url_for('watchlist.index'))
+
     return ui_blueprint

changedetectionio/blueprint/ui/diff.py

@@ -83,7 +83,6 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         If a processor doesn't have a difference module, falls back to text_json_diff.
         """
 
-        # More for testing, possible to return the first/only
         if uuid == 'first':
             uuid = list(datastore.data['watching'].keys()).pop()
 
@@ -101,23 +100,21 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         # Get the processor type for this watch
         processor_name = watch.get('processor', 'text_json_diff')
 
-        try:
-            # Try to import the processor's difference module
-            processor_module = importlib.import_module(f'changedetectionio.processors.{processor_name}.difference')
+        # Try to get the processor's difference module (works for both built-in and plugin processors)
+        from changedetectionio.processors import get_processor_submodule
+        processor_module = get_processor_submodule(processor_name, 'difference')
 
-            # Call the processor's render() function
-            if hasattr(processor_module, 'render'):
-                return processor_module.render(
-                    watch=watch,
-                    datastore=datastore,
-                    request=request,
-                    url_for=url_for,
-                    render_template=render_template,
-                    flash=flash,
-                    redirect=redirect
-                )
-        except (ImportError, ModuleNotFoundError) as e:
-            logger.warning(f"Processor {processor_name} does not have a difference module, falling back to text_json_diff: {e}")
+        # Call the processor's render() function
+        if processor_module and hasattr(processor_module, 'render'):
+            return processor_module.render(
+                watch=watch,
+                datastore=datastore,
+                request=request,
+                url_for=url_for,
+                render_template=render_template,
+                flash=flash,
+                redirect=redirect
+            )
 
         # Fallback: if processor doesn't have difference module, use text_json_diff as default
         from changedetectionio.processors.text_json_diff.difference import render as default_render
@@ -144,10 +141,10 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         Each processor implements processors/{type}/extract.py::render_form()
         If a processor doesn't have an extract module, falls back to text_json_diff.
         """
-        # More for testing, possible to return the first/only
+
+
         if uuid == 'first':
             uuid = list(datastore.data['watching'].keys()).pop()
-
         try:
             watch = datastore.data['watching'][uuid]
         except KeyError:
@@ -157,23 +154,21 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         # Get the processor type for this watch
         processor_name = watch.get('processor', 'text_json_diff')
 
-        try:
-            # Try to import the processor's extract module
-            processor_module = importlib.import_module(f'changedetectionio.processors.{processor_name}.extract')
+        # Try to get the processor's extract module (works for both built-in and plugin processors)
+        from changedetectionio.processors import get_processor_submodule
+        processor_module = get_processor_submodule(processor_name, 'extract')
 
-            # Call the processor's render_form() function
-            if hasattr(processor_module, 'render_form'):
-                return processor_module.render_form(
-                    watch=watch,
-                    datastore=datastore,
-                    request=request,
-                    url_for=url_for,
-                    render_template=render_template,
-                    flash=flash,
-                    redirect=redirect
-                )
-        except (ImportError, ModuleNotFoundError) as e:
-            logger.warning(f"Processor {processor_name} does not have an extract module, falling back to base extractor: {e}")
+        # Call the processor's render_form() function
+        if processor_module and hasattr(processor_module, 'render_form'):
+            return processor_module.render_form(
+                watch=watch,
+                datastore=datastore,
+                request=request,
+                url_for=url_for,
+                render_template=render_template,
+                flash=flash,
+                redirect=redirect
+            )
 
         # Fallback: if processor doesn't have extract module, use base processors.extract as default
         from changedetectionio.processors.extract import render_form as default_render_form
@@ -200,7 +195,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         Each processor implements processors/{type}/extract.py::process_extraction()
         If a processor doesn't have an extract module, falls back to text_json_diff.
         """
-        # More for testing, possible to return the first/only
+
         if uuid == 'first':
             uuid = list(datastore.data['watching'].keys()).pop()
 
@@ -213,24 +208,22 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         # Get the processor type for this watch
         processor_name = watch.get('processor', 'text_json_diff')
 
-        try:
-            # Try to import the processor's extract module
-            processor_module = importlib.import_module(f'changedetectionio.processors.{processor_name}.extract')
+        # Try to get the processor's extract module (works for both built-in and plugin processors)
+        from changedetectionio.processors import get_processor_submodule
+        processor_module = get_processor_submodule(processor_name, 'extract')
 
-            # Call the processor's process_extraction() function
-            if hasattr(processor_module, 'process_extraction'):
-                return processor_module.process_extraction(
-                    watch=watch,
-                    datastore=datastore,
-                    request=request,
-                    url_for=url_for,
-                    make_response=make_response,
-                    send_from_directory=send_from_directory,
-                    flash=flash,
-                    redirect=redirect
-                )
-        except (ImportError, ModuleNotFoundError) as e:
-            logger.warning(f"Processor {processor_name} does not have an extract module, falling back to base extractor: {e}")
+        # Call the processor's process_extraction() function
+        if processor_module and hasattr(processor_module, 'process_extraction'):
+            return processor_module.process_extraction(
+                watch=watch,
+                datastore=datastore,
+                request=request,
+                url_for=url_for,
+                make_response=make_response,
+                send_from_directory=send_from_directory,
+                flash=flash,
+                redirect=redirect
+            )
 
         # Fallback: if processor doesn't have extract module, use base processors.extract as default
         from changedetectionio.processors.extract import process_extraction as default_process_extraction
@@ -267,7 +260,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         - /diff/{uuid}/processor-asset/after
         - /diff/{uuid}/processor-asset/rendered_diff
         """
-        # More for testing, possible to return the first/only
+
         if uuid == 'first':
             uuid = list(datastore.data['watching'].keys()).pop()
 
@@ -280,38 +273,33 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         # Get the processor type for this watch
         processor_name = watch.get('processor', 'text_json_diff')
 
-        try:
-            # Try to import the processor's difference module
-            processor_module = importlib.import_module(f'changedetectionio.processors.{processor_name}.difference')
+        # Try to get the processor's difference module (works for both built-in and plugin processors)
+        from changedetectionio.processors import get_processor_submodule
+        processor_module = get_processor_submodule(processor_name, 'difference')
 
-            # Call the processor's get_asset() function
-            if hasattr(processor_module, 'get_asset'):
-                result = processor_module.get_asset(
-                    asset_name=asset_name,
-                    watch=watch,
-                    datastore=datastore,
-                    request=request
-                )
+        # Call the processor's get_asset() function
+        if processor_module and hasattr(processor_module, 'get_asset'):
+            result = processor_module.get_asset(
+                asset_name=asset_name,
+                watch=watch,
+                datastore=datastore,
+                request=request
+            )
 
-                if result is None:
-                    from flask import abort
-                    abort(404, description=f"Asset '{asset_name}' not found")
-
-                binary_data, content_type, cache_control = result
-
-                response = make_response(binary_data)
-                response.headers['Content-Type'] = content_type
-                if cache_control:
-                    response.headers['Cache-Control'] = cache_control
-                return response
-            else:
-                logger.warning(f"Processor {processor_name} does not implement get_asset()")
+            if result is None:
                 from flask import abort
-                abort(404, description=f"Processor '{processor_name}' does not support assets")
+                abort(404, description=f"Asset '{asset_name}' not found")
 
-        except (ImportError, ModuleNotFoundError) as e:
-            logger.warning(f"Processor {processor_name} does not have a difference module: {e}")
+            binary_data, content_type, cache_control = result
+
+            response = make_response(binary_data)
+            response.headers['Content-Type'] = content_type
+            if cache_control:
+                response.headers['Cache-Control'] = cache_control
+            return response
+        else:
+            logger.warning(f"Processor {processor_name} does not implement get_asset()")
             from flask import abort
-            abort(404, description=f"Processor '{processor_name}' not found")
+            abort(404, description=f"Processor '{processor_name}' does not support assets")
 
     return diff_blueprint

changedetectionio/blueprint/ui/edit.py

@@ -9,7 +9,7 @@ from jinja2 import Environment, FileSystemLoader
 from changedetectionio.store import ChangeDetectionStore
 from changedetectionio.auth_decorator import login_optionally_required
 from changedetectionio.time_handler import is_within_schedule
-from changedetectionio import worker_handler
+from changedetectionio import worker_pool
 
 def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMetaData):
     edit_blueprint = Blueprint('ui_edit', __name__, template_folder="../ui/templates")
@@ -30,14 +30,13 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
         from changedetectionio import processors
         import importlib
 
+        if uuid == 'first':
+            uuid = list(datastore.data['watching'].keys()).pop()
         # More for testing, possible to return the first/only
         if not datastore.data['watching'].keys():
             flash(gettext("No watches to edit"), "error")
             return redirect(url_for('watchlist.index'))
 
-        if uuid == 'first':
-            uuid = list(datastore.data['watching'].keys()).pop()
-
         if not uuid in datastore.data['watching']:
             flash(gettext("No watch with the UUID {} found.").format(uuid), "error")
             return redirect(url_for('watchlist.index'))
@@ -72,8 +71,13 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
         processor_name = datastore.data['watching'][uuid].get('processor', '')
         processor_classes = next((tpl for tpl in processors.find_processors() if tpl[1] == processor_name), None)
         if not processor_classes:
-            flash(gettext("Cannot load the edit form for processor/plugin '{}', plugin missing?").format(processor_classes[1]), 'error')
-            return redirect(url_for('watchlist.index'))
+            flash(gettext("Could not load '{}' processor, processor plugin might be missing. Please select a different processor.").format(processor_name), 'error')
+            # Fall back to default processor so user can still edit and change processor
+            processor_classes = next((tpl for tpl in processors.find_processors() if tpl[1] == 'text_json_diff'), None)
+            if not processor_classes:
+                # If even text_json_diff is missing, something is very wrong
+                flash(gettext("Could not load '{}' processor, processor plugin might be missing.").format(processor_name), 'error')
+                return redirect(url_for('watchlist.index'))
 
         parent_module = processors.get_parent_module(processor_classes[0])
 
@@ -150,58 +154,10 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
 
             extra_update_obj['time_between_check'] = form.time_between_check.data
 
-             # Handle processor-config-* fields separately (save to JSON, not datastore)
-            processor_config_data = {}
-            fields_to_remove = []
-            for field_name, field_value in form.data.items():
-                if field_name.startswith('processor_config_'):
-                    config_key = field_name.replace('processor_config_', '')
-                    if field_value:  # Only save non-empty values
-                        processor_config_data[config_key] = field_value
-                    fields_to_remove.append(field_name)
-
-            # Save processor config to JSON file if any config data exists
-            if processor_config_data:
-                try:
-                    processor_name = form.data.get('processor')
-                    # Create a processor instance to access config methods
-                    processor_instance = processors.difference_detection_processor(datastore, uuid)
-                    # Use processor name as filename so each processor keeps its own config
-                    config_filename = f'{processor_name}.json'
-                    processor_instance.update_extra_watch_config(config_filename, processor_config_data)
-                    logger.debug(f"Saved processor config to {config_filename}: {processor_config_data}")
-
-                    # Call optional edit_hook if processor has one
-                    try:
-                        # Try to import the edit_hook module from the processor package
-                        import importlib
-                        edit_hook_module_name = f'changedetectionio.processors.{processor_name}.edit_hook'
-
-                        try:
-                            edit_hook = importlib.import_module(edit_hook_module_name)
-                            logger.debug(f"Found edit_hook module for {processor_name}")
-
-                            if hasattr(edit_hook, 'on_config_save'):
-                                logger.info(f"Calling edit_hook.on_config_save for {processor_name}")
-                                watch_obj = datastore.data['watching'][uuid]
-                                # Call hook and get updated config
-                                updated_config = edit_hook.on_config_save(watch_obj, processor_config_data, datastore)
-                                # Save updated config back to file
-                                processor_instance.update_extra_watch_config(config_filename, updated_config)
-                                logger.info(f"Edit hook updated config: {updated_config}")
-                            else:
-                                logger.debug(f"Edit hook module found but no on_config_save function")
-                        except ModuleNotFoundError:
-                            logger.debug(f"No edit_hook module for processor {processor_name} (this is normal)")
-                    except Exception as hook_error:
-                        logger.error(f"Edit hook error (non-fatal): {hook_error}", exc_info=True)
-
-                except Exception as e:
-                    logger.error(f"Failed to save processor config: {e}")
-
-            # Remove processor-config-* fields from form.data before updating datastore
-            for field_name in fields_to_remove:
-                form.data.pop(field_name, None)
+            # Handle processor-config-* fields separately (save to JSON, not datastore)
+            # IMPORTANT: These must NOT be saved to url-watches.json, only to the processor-specific JSON file
+            processor_config_data = processors.extract_processor_config_from_form_data(form.data)
+            processors.save_processor_config(datastore, uuid, processor_config_data)
 
             # Ignore text
             form_ignore_text = form.ignore_text.data
@@ -241,7 +197,11 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
 
             # Recast it if need be to right data Watch handler
             watch_class = processors.get_custom_watch_obj_for_processor(form.data.get('processor'))
-            datastore.data['watching'][uuid] = watch_class(datastore_path=datastore.datastore_path, default=datastore.data['watching'][uuid])
+            datastore.data['watching'][uuid] = watch_class(datastore_path=datastore.datastore_path, __datastore=datastore.data, default=datastore.data['watching'][uuid])
+
+            # Save the watch immediately
+            datastore.data['watching'][uuid].commit()
+
             flash(gettext("Updated watch - unpaused!") if request.args.get('unpause_on_save') else gettext("Updated watch."))
 
             # Cleanup any browsersteps session for this watch
@@ -251,10 +211,6 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
             except Exception as e:
                 logger.debug(f"Error cleaning up browsersteps session: {e}")
 
-            # Re #286 - We wait for syncing new data to disk in another thread every 60 seconds
-            # But in the case something is added we should save straight away
-            datastore.needs_write_urgent = True
-
             # Do not queue on edit if its not within the time range
 
             # @todo maybe it should never queue anyway on edit...
@@ -283,7 +239,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
             #############################
             if not datastore.data['watching'][uuid].get('paused') and is_in_schedule:
                 # Queue the watch for immediate recheck, with a higher priority
-                worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+                worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
 
             # Diff page [edit] link should go back to diff page
             if request.args.get("next") and request.args.get("next") == 'diff':
@@ -311,10 +267,17 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
 
             # Get fetcher capabilities instead of hardcoded logic
             capabilities = get_fetcher_capabilities(watch, datastore)
+
+            # Add processor capabilities from module
+            capabilities['supports_visual_selector'] = getattr(parent_module, 'supports_visual_selector', False)
+            capabilities['supports_text_filters_and_triggers'] = getattr(parent_module, 'supports_text_filters_and_triggers', False)
+            capabilities['supports_text_filters_and_triggers_elements'] = getattr(parent_module, 'supports_text_filters_and_triggers_elements', False)
+            capabilities['supports_request_type'] = getattr(parent_module, 'supports_request_type', False)
+
             app_rss_token = datastore.data['settings']['application'].get('rss_access_token'),
 
             c = [f"processor-{watch.get('processor')}"]
-            if worker_handler.is_watch_running(uuid):
+            if worker_pool.is_watch_running(uuid):
                 c.append('checking-now')
 
             template_args = {
@@ -371,10 +334,12 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
         from flask import send_file
         import brotli
 
+        if uuid == 'first':
+            uuid = list(datastore.data['watching'].keys()).pop()
         watch = datastore.data['watching'].get(uuid)
-        if watch and watch.history.keys() and os.path.isdir(watch.watch_data_dir):
+        if watch and watch.history.keys() and os.path.isdir(watch.data_dir):
             latest_filename = list(watch.history.keys())[-1]
-            html_fname = os.path.join(watch.watch_data_dir, f"{latest_filename}.html.br")
+            html_fname = os.path.join(watch.data_dir, f"{latest_filename}.html.br")
             with open(html_fname, 'rb') as f:
                 if html_fname.endswith('.br'):
                     # Read and decompress the Brotli file
@@ -395,6 +360,9 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
     def watch_get_preview_rendered(uuid):
         '''For when viewing the "preview" of the rendered text from inside of Edit'''
         from flask import jsonify
+
+        if uuid == 'first':
+            uuid = list(datastore.data['watching'].keys()).pop()
         from changedetectionio.processors.text_json_diff import prepare_filter_prevew
         result = prepare_filter_prevew(watch_uuid=uuid, form_data=request.form, datastore=datastore)
         return jsonify(result)
@@ -418,6 +386,9 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
                     s = re.sub(r'[0-9]+', r'\\d+', s)
                     datastore.data["watching"][uuid]['ignore_text'].append('/' + s + '/')
 
+            # Save the updated ignore_text
+            datastore.data["watching"][uuid].commit()
+
         return f"<a href={url_for('ui.ui_preview.preview_page', uuid=uuid)}>Click to preview</a>"
     
     return edit_blueprint

changedetectionio/blueprint/ui/preview.py

@@ -26,10 +26,9 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         Each processor implements processors/{type}/preview.py::render()
         If a processor doesn't have a preview module, falls back to default text preview.
         """
-        # More for testing, possible to return the first/only
+
         if uuid == 'first':
             uuid = list(datastore.data['watching'].keys()).pop()
-
         try:
             watch = datastore.data['watching'][uuid]
         except KeyError:
@@ -39,24 +38,21 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         # Get the processor type for this watch
         processor_name = watch.get('processor', 'text_json_diff')
 
-        try:
-            # Try to import the processor's preview module
-            import importlib
-            processor_module = importlib.import_module(f'changedetectionio.processors.{processor_name}.preview')
+        # Try to get the processor's preview module (works for both built-in and plugin processors)
+        from changedetectionio.processors import get_processor_submodule
+        processor_module = get_processor_submodule(processor_name, 'preview')
 
-            # Call the processor's render() function
-            if hasattr(processor_module, 'render'):
-                return processor_module.render(
-                    watch=watch,
-                    datastore=datastore,
-                    request=request,
-                    url_for=url_for,
-                    render_template=render_template,
-                    flash=flash,
-                    redirect=redirect
-                )
-        except (ImportError, ModuleNotFoundError) as e:
-            logger.debug(f"Processor {processor_name} does not have a preview module, using default preview: {e}")
+        # Call the processor's render() function
+        if processor_module and hasattr(processor_module, 'render'):
+            return processor_module.render(
+                watch=watch,
+                datastore=datastore,
+                request=request,
+                url_for=url_for,
+                render_template=render_template,
+                flash=flash,
+                redirect=redirect
+            )
 
         # Fallback: if processor doesn't have preview module, use default text preview
         content = []
@@ -150,10 +146,8 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         """
         from flask import make_response
 
-        # More for testing, possible to return the first/only
         if uuid == 'first':
             uuid = list(datastore.data['watching'].keys()).pop()
-
         try:
             watch = datastore.data['watching'][uuid]
         except KeyError:
@@ -163,39 +157,33 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         # Get the processor type for this watch
         processor_name = watch.get('processor', 'text_json_diff')
 
-        try:
-            # Try to import the processor's preview module
-            import importlib
-            processor_module = importlib.import_module(f'changedetectionio.processors.{processor_name}.preview')
+        # Try to get the processor's preview module (works for both built-in and plugin processors)
+        from changedetectionio.processors import get_processor_submodule
+        processor_module = get_processor_submodule(processor_name, 'preview')
 
-            # Call the processor's get_asset() function
-            if hasattr(processor_module, 'get_asset'):
-                result = processor_module.get_asset(
-                    asset_name=asset_name,
-                    watch=watch,
-                    datastore=datastore,
-                    request=request
-                )
+        # Call the processor's get_asset() function
+        if processor_module and hasattr(processor_module, 'get_asset'):
+            result = processor_module.get_asset(
+                asset_name=asset_name,
+                watch=watch,
+                datastore=datastore,
+                request=request
+            )
 
-                if result is None:
-                    from flask import abort
-                    abort(404, description=f"Asset '{asset_name}' not found")
-
-                binary_data, content_type, cache_control = result
-
-                response = make_response(binary_data)
-                response.headers['Content-Type'] = content_type
-                if cache_control:
-                    response.headers['Cache-Control'] = cache_control
-                return response
-            else:
-                logger.warning(f"Processor {processor_name} does not implement get_asset()")
+            if result is None:
                 from flask import abort
-                abort(404, description=f"Processor '{processor_name}' does not support assets")
+                abort(404, description=f"Asset '{asset_name}' not found")
 
-        except (ImportError, ModuleNotFoundError) as e:
-            logger.warning(f"Processor {processor_name} does not have a preview module: {e}")
+            binary_data, content_type, cache_control = result
+
+            response = make_response(binary_data)
+            response.headers['Content-Type'] = content_type
+            if cache_control:
+                response.headers['Cache-Control'] = cache_control
+            return response
+        else:
+            logger.warning(f"Processor {processor_name} does not implement get_asset()")
             from flask import abort
-            abort(404, description=f"Processor '{processor_name}' not found")
+            abort(404, description=f"Processor '{processor_name}' does not support assets")
 
     return preview_blueprint

changedetectionio/blueprint/ui/templates/edit.html

@@ -45,14 +45,19 @@
     <div class="tabs collapsable">
         <ul>
             <li class="tab"><a href="#general">{{ _('General') }}</a></li>
+            {% if capabilities.supports_request_type %}
             <li class="tab"><a href="#request">{{ _('Request') }}</a></li>
+            {% endif %}
             {% if extra_tab_content %}
             <li class="tab"><a href="#extras_tab">{{ extra_tab_content }}</a></li>
             {% endif %}
+            {% if capabilities.supports_browser_steps %}
             <li class="tab"><a id="browsersteps-tab" href="#browser-steps">{{ _('Browser Steps') }}</a></li>
-        <!-- should goto extra forms? -->
-            {% if watch['processor'] == 'text_json_diff' or watch['processor'] == 'image_ssim_diff' %}
+            {% endif %}
+            {% if capabilities.supports_visual_selector %}
             <li class="tab"><a id="visualselector-tab" href="#visualselector">{{ _('Visual Filter Selector') }}</a></li>
+            {% endif %}
+            {% if capabilities.supports_text_filters_and_triggers %}
             <li class="tab" id="filters-and-triggers-tab"><a href="#filters-and-triggers">{{ _('Filters & Triggers') }}</a></li>
             <li class="tab" id="conditions-tab"><a href="#conditions">{{ _('Conditions') }}</a></li>
             {% endif %}
@@ -110,12 +115,20 @@
                          {{ _('Sends a notification when the filter can no longer be seen on the page, good for knowing when the page changed and your filter will not work anymore.') }}
                         </span>
                     </div>
+                    <div class="pure-control-group">
+                        {{ render_field(form.history_snapshot_max_length, class="history_snapshot_max_length") }}
+                        <span class="pure-form-message-inline">{{ _('Limit collection of history snapshots for each watch to this number of history items.') }}
+                        <br>
+                        {{ _('Set to empty to use system settings default') }}
+                        </span>
+                    </div>
                     <div class="pure-control-group">
                         {{ render_ternary_field(form.use_page_title_in_list) }}
                     </div>
                 </fieldset>
             </div>
 
+            {% if capabilities.supports_request_type %}
             <div class="tab-pane-inner" id="request">
                     <div class="pure-control-group inline-radio">
                         {{ render_field(form.fetch_backend, class="fetch-backend") }}
@@ -203,6 +216,7 @@ Math: {{ 1 + 1 }}") }}
                     </div>
             </fieldset>
             </div>
+            {% endif %}
 
             <div class="tab-pane-inner" id="browser-steps">
             {% if capabilities.supports_browser_steps %}
@@ -283,8 +297,7 @@ Math: {{ 1 + 1 }}") }}
                 </fieldset>
             </div>
 
-            {% if watch['processor'] == 'text_json_diff' or watch['processor'] == 'image_ssim_diff' %}
-
+            {% if capabilities.supports_text_filters_and_triggers %}
             <div class="tab-pane-inner" id="conditions">
                     <script>
                         const verify_condition_rule_url="{{url_for('conditions.verify_condition_single_rule', watch_uuid=uuid)}}";
@@ -303,7 +316,9 @@ Math: {{ 1 + 1 }}") }}
                 <span id="activate-text-preview" class="pure-button pure-button-primary button-xsmall">{{ _('Activate preview') }}</span>
               <div>
               <div id="edit-text-filter">
-                    <div class="pure-control-group" id="pro-tips">
+
+{% if capabilities.supports_text_filters_and_triggers_elements %}
+                        <div class="pure-control-group" id="pro-tips">
                             <strong>{{ _('Pro-tips:') }}</strong><br>
                             <ul>
                                 <li>
@@ -314,8 +329,8 @@ Math: {{ 1 + 1 }}") }}
                                 </li>
                             </ul>
                     </div>
-
 {% include "edit/include_subtract.html" %}
+{% endif %}
                 <div class="text-filtering border-fieldset">
                 <fieldset class="pure-group" id="text-filtering-type-options">
                     <h3>{{ _('Text filtering') }}</h3>
@@ -374,7 +389,7 @@ Math: {{ 1 + 1 }}") }}
             {{ extra_form_content|safe }}
             </div>
         {% endif %}
-            {% if watch['processor'] == 'text_json_diff' or watch['processor'] == 'image_ssim_diff' %}
+            {% if capabilities.supports_visual_selector %}
             <div class="tab-pane-inner visual-selector-ui" id="visualselector">
                 <img class="beta-logo" src="{{url_for('static_content', group='images', filename='beta-logo.png')}}" alt="New beta functionality">
 
@@ -386,7 +401,7 @@ Math: {{ 1 + 1 }}") }}
                                     {{ _('The Visual Selector tool lets you select the') }} <i>{{ _('text') }}</i> {{ _('elements that will be used for the change detection. It automatically fills-in the filters in the "CSS/JSONPath/JQ/XPath Filters" box of the') }} <a href="#filters-and-triggers">{{ _('Filters & Triggers') }}</a> {{ _('tab. Use') }} <strong>{{ _('Shift+Click') }}</strong> {{ _('to select multiple items.') }}
                                 </span>
 
-                                {% if watch['processor'] == 'image_ssim_diff' %}
+                                {% if watch['processor'] == 'image_ssim_diff' %} {# @todo, integrate with image_ssim_diff selector better, use some extra form ? #}
                                 <div id="selection-mode-controls" style="margin: 10px 0; padding: 10px; background: var(--color-background-tab); border-radius: 5px;">
                                     <label style="font-weight: 600; margin-right: 15px;">{{ _('Selection Mode:') }}</label>
                                     <label style="margin-right: 15px;">

changedetectionio/blueprint/ui/views.py

@@ -2,7 +2,7 @@ from flask import Blueprint, request, redirect, url_for, flash
 from flask_babel import gettext
 from changedetectionio.store import ChangeDetectionStore
 from changedetectionio.auth_decorator import login_optionally_required
-from changedetectionio import worker_handler
+from changedetectionio import worker_pool
 
 
 def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMetaData, watch_check_update):
@@ -24,8 +24,9 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
             flash(gettext('Warning, URL {} already exists').format(url), "notice")
 
         add_paused = request.form.get('edit_and_watch_submit_button') != None
-        processor = request.form.get('processor', 'text_json_diff')
-        new_uuid = datastore.add_watch(url=url, tag=request.form.get('tags').strip(), extras={'paused': add_paused, 'processor': processor})
+        from changedetectionio import processors
+        processor = request.form.get('processor', processors.get_default_processor())
+        new_uuid = datastore.add_watch(url=url, tag=request.form.get('tags','').strip(), extras={'paused': add_paused, 'processor': processor})
 
         if new_uuid:
             if add_paused:
@@ -33,9 +34,9 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
                 return redirect(url_for('ui.ui_edit.edit_page', uuid=new_uuid, unpause_on_save=1, tag=request.args.get('tag')))
             else:
                 # Straight into the queue.
-                worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': new_uuid}))
+                worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': new_uuid}))
                 flash(gettext("Watch added."))
 
         return redirect(url_for('watchlist.index', tag=request.args.get('tag','')))
 
-    return views_blueprint
+    return views_blueprint

changedetectionio/blueprint/watchlist/__init__.py

@@ -39,7 +39,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
             elif op == 'mute':
                 datastore.data['watching'][uuid].toggle_mute()
 
-            datastore.needs_write = True
+            datastore.data['watching'][uuid].commit()
             return redirect(url_for('watchlist.index', tag = active_tag_uuid))
 
         # Sort by last_changed and add the uuid which is usually the key..

changedetectionio/blueprint/watchlist/templates/watch-overview.html

@@ -1,5 +1,9 @@
 {%- extends 'base.html' -%}
 {%- block content -%}
+{%- set tips = [
+    _("Changedetection.io can monitor more than just web-pages! See our plugins!") ~ ' <a href="https://changedetection.io/plugins">' ~ _('More info') ~ '</a>',
+    _("You can also add 'shared' watches.") ~ ' <a href="https://github.com/dgtlmoon/changedetection.io/wiki/Sharing-a-Watch">' ~ _('More info') ~ '</a>'
+] -%}
 {%- from '_helpers.html' import render_simple_field, render_field, render_nolabel_field, sort_by_title -%}
 <script src="{{url_for('static_content', group='js', filename='jquery-3.6.0.min.js')}}"></script>
 <script src="{{url_for('static_content', group='js', filename='watch-overview.js')}}" defer></script>
@@ -10,6 +14,46 @@
 // Initialize Feather icons after the page loads
 document.addEventListener('DOMContentLoaded', function() {
     feather.replace();
+
+    // Intersection Observer for lazy loading favicons
+    // Only load favicon images when they enter the viewport
+    if ('IntersectionObserver' in window) {
+        const faviconObserver = new IntersectionObserver((entries, observer) => {
+            entries.forEach(entry => {
+                if (entry.isIntersecting) {
+                    const img = entry.target;
+                    const src = img.getAttribute('data-src');
+
+                    if (src) {
+                        // Load the actual favicon
+                        img.src = src;
+                        img.removeAttribute('data-src');
+                    }
+
+                    // Stop observing this image
+                    observer.unobserve(img);
+                }
+            });
+        }, {
+            // Start loading slightly before the image enters viewport
+            rootMargin: '50px',
+            threshold: 0.01
+        });
+
+        // Observe all lazy favicon images
+        document.querySelectorAll('.lazy-favicon').forEach(img => {
+            faviconObserver.observe(img);
+        });
+    } else {
+        // Fallback for older browsers: load all favicons immediately
+        document.querySelectorAll('.lazy-favicon').forEach(img => {
+            const src = img.getAttribute('data-src');
+            if (src) {
+                img.src = src;
+                img.removeAttribute('data-src');
+            }
+        });
+    }
 });
 </script>
 <style>
@@ -69,7 +113,9 @@ html[data-darkmode="true"] .watch-tag-list.tag-{{ class_name }} {
             </div>
 
         </fieldset>
-        <span style="color:#eee; font-size: 80%;"><img alt="{{ _('Create a shareable link') }}" style="height: 1em;display:inline-block;" src="{{url_for('static_content', group='images', filename='spread-white.svg')}}" > {{ _("Tip: You can also add 'shared' watches.") }} <a href="https://github.com/dgtlmoon/changedetection.io/wiki/Sharing-a-Watch">{{ _('More info') }}</a></span>
+        <span style="color:#eee; font-size: 80%;">
+            <strong>Tip: </strong> {{ tips | random | safe }}
+        </span>
     </form>
 </div>
 <div class="box">
@@ -200,30 +246,40 @@ html[data-darkmode="true"] .watch-tag-list.tag-{{ class_name }} {
                 <td class="title-col inline">
                     <div class="flex-wrapper">
                     {% if 'favicons_enabled' not in ui_settings or ui_settings['favicons_enabled'] %}
-                        <div>{# A page might have hundreds of these images, set IMG options for lazy loading, don't set SRC if we dont have it so it doesnt fetch the placeholder'  #}
-                            <img alt="Favicon thumbnail" class="favicon" loading="lazy" decoding="async" fetchpriority="low" {% if favicon %} src="{{url_for('static_content', group='favicon', filename=watch.uuid)}}" {% else %} src='data:image/svg+xml;utf8,%3Csvg xmlns="http://www.w3.org/2000/svg" width="7.087" height="7.087" viewBox="0 0 7.087 7.087"%3E%3Ccircle cx="3.543" cy="3.543" r="3.279" stroke="%23e1e1e1" stroke-width="0.45" fill="none" opacity="0.74"/%3E%3C/svg%3E' {%  endif %} >
+                        <div>
+                            {# Intersection Observer lazy loading: store real URL in data-src, load only when visible in viewport #}
+                            <img alt="Favicon thumbnail"
+                                 class="favicon lazy-favicon"
+                                 loading="lazy"
+                                 decoding="async"
+                                 fetchpriority="low"
+                                 {% if favicon %}
+                                 data-src="{{url_for('static_content', group='favicon', filename=watch.uuid)}}"
+                                 {% endif %}
+                                 src='data:image/svg+xml;utf8,%3Csvg xmlns="http://www.w3.org/2000/svg" width="7.087" height="7.087" viewBox="0 0 7.087 7.087"%3E%3Ccircle cx="3.543" cy="3.543" r="3.279" stroke="%23e1e1e1" stroke-width="0.45" fill="none" opacity="0.74"/%3E%3C/svg%3E'>
                         </div>
                     {%  endif %}
                         <div>
-                        <span class="watch-title">
-                            {% if system_use_url_watchlist or watch.get('use_page_title_in_list') %}
-                                {{ watch.label }}
-                            {% else %}
-                                {{ watch.get('title') or watch.link }}
-                            {% endif %}
-                           <a class="external" target="_blank" rel="noopener" href="{{ watch.link.replace('source:','') }}">&nbsp;</a>
-                        </span>
+                            {%- if watch['processor'] and watch['processor'] in processor_badge_texts -%}
+                                <span class="processor-badge processor-badge-{{ watch['processor'] }}" title="{{ processor_descriptions.get(watch['processor'], watch['processor']) }}">{{ processor_badge_texts[watch['processor']] }}</span>
+                            {%- endif -%}
+                            <span class="watch-title">
+                                {% if system_use_url_watchlist or watch.get('use_page_title_in_list') %}
+                                    {{ watch.label }}
+                                {% else %}
+                                    {{ watch.get('title') or watch.link }}
+                                {% endif %}
+                               <a class="external" target="_blank" rel="noopener" href="{{ watch.link.replace('source:','') }}">&nbsp;</a>
+                            </span>
                             <div class="error-text" style="display:none;">{{ watch.compile_error_texts(has_proxies=datastore.proxy_list)|safe }}</div>
                             {%- if watch['processor'] == 'text_json_diff'  -%}
                                 {%- if watch['has_ldjson_price_data'] and not watch['track_ldjson_price_data']  -%}
                                 <div class="ldjson-price-track-offer">Switch to Restock & Price watch mode? <a href="{{url_for('price_data_follower.accept', uuid=watch.uuid)}}" class="pure-button button-xsmall">Yes</a> <a href="{{url_for('price_data_follower.reject', uuid=watch.uuid)}}" class="">No</a></div>
                                 {%- endif -%}
                             {%- endif -%}
-                            {%- if watch['processor'] and watch['processor'] in processor_badge_texts -%}
-                                <span class="processor-badge processor-badge-{{ watch['processor'] }}" title="{{ processor_descriptions.get(watch['processor'], watch['processor']) }}">{{ processor_badge_texts[watch['processor']] }}</span>
-                            {%- endif -%}
+
                             {%- for watch_tag_uuid, watch_tag in datastore.get_all_tags_for_watch(watch['uuid']).items() -%}
-                              <span class="watch-tag-list tag-{{ watch_tag.title|sanitize_tag_class }}">{{ watch_tag.title }}</span>
+                              <a href="{{url_for('watchlist.index', tag=watch_tag_uuid) }}" class="watch-tag-list tag-{{ watch_tag.title|sanitize_tag_class }}">{{ watch_tag.title }}</a>
                             {%- endfor -%}
                         </div>
                     <div class="status-icons">

changedetectionio/content_fetchers/base.py

@@ -71,10 +71,19 @@ class Fetcher():
     supports_screenshots = False        # Can capture page screenshots
     supports_xpath_element_data = False # Can extract xpath element positions/data for visual selector
 
+    # Screenshot element locking - prevents layout shifts during screenshot capture
+    # Only needed for visual comparison (image_ssim_diff processor)
+    # Locks element dimensions in the first viewport to prevent headers/ads from resizing
+    lock_viewport_elements = False      # Default: disabled for performance
+
     def __init__(self, **kwargs):
         if kwargs and 'screenshot_format' in kwargs:
             self.screenshot_format = kwargs.get('screenshot_format')
 
+        # Allow lock_viewport_elements to be set via kwargs
+        if kwargs and 'lock_viewport_elements' in kwargs:
+            self.lock_viewport_elements = kwargs.get('lock_viewport_elements')
+
 
     @classmethod
     def get_status_icon_data(cls):

changedetectionio/content_fetchers/playwright.py

@@ -1,3 +1,4 @@
+import asyncio
 import gc
 import json
 import os
@@ -8,20 +9,24 @@ from loguru import logger
 from changedetectionio.content_fetchers import SCREENSHOT_MAX_HEIGHT_DEFAULT, visualselector_xpath_selectors, \
     SCREENSHOT_SIZE_STITCH_THRESHOLD, SCREENSHOT_MAX_TOTAL_HEIGHT, XPATH_ELEMENT_JS, INSTOCK_DATA_JS, FAVICON_FETCHER_JS
 from changedetectionio.content_fetchers.base import Fetcher, manage_user_agent
-from changedetectionio.content_fetchers.exceptions import PageUnloadable, Non200ErrorCodeReceived, EmptyReply, ScreenshotUnavailable
+from changedetectionio.content_fetchers.exceptions import PageUnloadable, Non200ErrorCodeReceived, EmptyReply, ScreenshotUnavailable, \
+    BrowserStepsStepException
 
-async def capture_full_page_async(page, screenshot_format='JPEG'):
+
+async def capture_full_page_async(page, screenshot_format='JPEG', watch_uuid=None, lock_viewport_elements=False):
     import os
     import time
-    import multiprocessing
 
     start = time.time()
+    watch_info = f"[{watch_uuid}] " if watch_uuid else ""
 
+    setup_start = time.time()
     page_height = await page.evaluate("document.documentElement.scrollHeight")
     page_width = await page.evaluate("document.documentElement.scrollWidth")
     original_viewport = page.viewport_size
+    dimensions_time = time.time() - setup_start
 
-    logger.debug(f"Playwright viewport size {page.viewport_size} page height {page_height} page width {page_width}")
+    logger.debug(f"{watch_info}Playwright viewport size {page.viewport_size} page height {page_height} page width {page_width} (got dimensions in {dimensions_time:.2f}s)")
 
     # Use an approach similar to puppeteer: set a larger viewport and take screenshots in chunks
     step_size = SCREENSHOT_SIZE_STITCH_THRESHOLD # Size that won't cause GPU to overflow
@@ -29,25 +34,31 @@ async def capture_full_page_async(page, screenshot_format='JPEG'):
     y = 0
     elements_locked = False
 
-    if page_height > page.viewport_size['height']:
-
-        # Lock all element dimensions BEFORE screenshot to prevent CSS media queries from resizing
-        # capture_full_page_async() changes viewport height which triggers @media (min-height) rules
+    # Only lock viewport elements if explicitly enabled (for image_ssim_diff processor)
+    # This prevents headers/ads from resizing when viewport changes
+    if lock_viewport_elements and page_height > page.viewport_size['height']:
+        lock_start = time.time()
         lock_elements_js_path = os.path.join(os.path.dirname(__file__), 'res', 'lock-elements-sizing.js')
         with open(lock_elements_js_path, 'r') as f:
             lock_elements_js = f.read()
         await page.evaluate(lock_elements_js)
         elements_locked = True
+        lock_time = time.time() - lock_start
+        logger.debug(f"{watch_info}Viewport element locking enabled (took {lock_time:.2f}s)")
 
-        logger.debug("Element dimensions locked before screenshot capture")
-
+    if page_height > page.viewport_size['height']:
         if page_height < step_size:
             step_size = page_height # Incase page is bigger than default viewport but smaller than proposed step size
-        logger.debug(f"Setting bigger viewport to step through large page width W{page.viewport_size['width']}xH{step_size} because page_height > viewport_size")
+        viewport_start = time.time()
+        logger.debug(f"{watch_info}Setting bigger viewport to step through large page width W{page.viewport_size['width']}xH{step_size} because page_height > viewport_size")
         # Set viewport to a larger size to capture more content at once
         await page.set_viewport_size({'width': page.viewport_size['width'], 'height': step_size})
+        viewport_time = time.time() - viewport_start
+        logger.debug(f"{watch_info}Viewport changed to {page.viewport_size['width']}x{step_size} (took {viewport_time:.2f}s)")
 
     # Capture screenshots in chunks up to the max total height
+    capture_start = time.time()
+    chunk_times = []
     # Use PNG for better quality (no compression artifacts), JPEG for smaller size
     screenshot_type = screenshot_format.lower() if screenshot_format else 'jpeg'
     # PNG should use quality 100, JPEG uses configurable quality
@@ -69,7 +80,11 @@ async def capture_full_page_async(page, screenshot_format='JPEG'):
         if screenshot_type == 'jpeg':
             screenshot_kwargs['quality'] = screenshot_quality
 
+        chunk_start = time.time()
         screenshot_chunks.append(await page.screenshot(**screenshot_kwargs))
+        chunk_time = time.time() - chunk_start
+        chunk_times.append(chunk_time)
+        logger.debug(f"{watch_info}Chunk {len(screenshot_chunks)} captured in {chunk_time:.2f}s")
         y += step_size
 
     # Restore original viewport size
@@ -81,40 +96,54 @@ async def capture_full_page_async(page, screenshot_format='JPEG'):
         with open(unlock_elements_js_path, 'r') as f:
             unlock_elements_js = f.read()
         await page.evaluate(unlock_elements_js)
-        logger.debug("Element dimensions unlocked after screenshot capture")
+        logger.debug(f"{watch_info}Element dimensions unlocked after screenshot capture")
+
+    capture_time = time.time() - capture_start
+    total_capture_time = sum(chunk_times)
+    logger.debug(f"{watch_info}All {len(screenshot_chunks)} chunks captured in {capture_time:.2f}s (total chunk time: {total_capture_time:.2f}s)")
 
     # If we have multiple chunks, stitch them together
     if len(screenshot_chunks) > 1:
-        logger.debug(f"Screenshot stitching {len(screenshot_chunks)} chunks together")
+        stitch_start = time.time()
+        logger.debug(f"{watch_info}Starting stitching of {len(screenshot_chunks)} chunks")
 
-        # For small number of chunks (2-3), stitch inline to avoid multiprocessing overhead
-        # Only use separate process for many chunks (4+) to avoid blocking the event loop
-        if len(screenshot_chunks) <= 3:
-            from changedetectionio.content_fetchers.screenshot_handler import stitch_images_inline
-            screenshot = stitch_images_inline(screenshot_chunks, page_height, SCREENSHOT_MAX_TOTAL_HEIGHT)
-        else:
-            # Use separate process for many chunks to avoid blocking
-            # Always use spawn for thread safety - consistent behavior in tests and production
-            from changedetectionio.content_fetchers.screenshot_handler import stitch_images_worker
-            ctx = multiprocessing.get_context('spawn')
-            parent_conn, child_conn = ctx.Pipe()
-            p = ctx.Process(target=stitch_images_worker, args=(child_conn, screenshot_chunks, page_height, SCREENSHOT_MAX_TOTAL_HEIGHT))
-            p.start()
-            screenshot = parent_conn.recv_bytes()
-            p.join()
-            # Explicit cleanup
-            del p
-            del parent_conn, child_conn
+        # Always use spawn subprocess for ANY stitching (2+ chunks)
+        # PIL allocates at C level and Python GC never releases it - subprocess exit forces OS to reclaim
+        # Trade-off: 35MB resource_tracker vs 500MB+ PIL leak in main process
+        from changedetectionio.content_fetchers.screenshot_handler import stitch_images_worker_raw_bytes
+        import multiprocessing
+        import struct
 
+        ctx = multiprocessing.get_context('spawn')
+        parent_conn, child_conn = ctx.Pipe()
+        p = ctx.Process(target=stitch_images_worker_raw_bytes, args=(child_conn, page_height, SCREENSHOT_MAX_TOTAL_HEIGHT))
+        p.start()
+
+        # Send via raw bytes (no pickle)
+        parent_conn.send_bytes(struct.pack('I', len(screenshot_chunks)))
+        for chunk in screenshot_chunks:
+            parent_conn.send_bytes(chunk)
+
+        screenshot = parent_conn.recv_bytes()
+        p.join()
+
+        parent_conn.close()
+        child_conn.close()
+        del p, parent_conn, child_conn
+
+        stitch_time = time.time() - stitch_start
+        total_time = time.time() - start
+        setup_time = total_time - capture_time - stitch_time
         logger.debug(
-            f"Screenshot (chunked/stitched) - Page height: {page_height} Capture height: {SCREENSHOT_MAX_TOTAL_HEIGHT} - Stitched together in {time.time() - start:.2f}s")
-        # Explicit cleanup
-        del screenshot_chunks
-        screenshot_chunks = None
+            f"{watch_info}Screenshot complete - Page height: {page_height}px, Capture height: {SCREENSHOT_MAX_TOTAL_HEIGHT}px | "
+            f"Setup: {setup_time:.2f}s, Capture: {capture_time:.2f}s, Stitching: {stitch_time:.2f}s, Total: {total_time:.2f}s")
         return screenshot
 
+    total_time = time.time() - start
+    setup_time = total_time - capture_time
     logger.debug(
-        f"Screenshot Page height: {page_height} Capture height: {SCREENSHOT_MAX_TOTAL_HEIGHT} - Stitched together in {time.time() - start:.2f}s")
+        f"{watch_info}Screenshot complete - Page height: {page_height}px, Capture height: {SCREENSHOT_MAX_TOTAL_HEIGHT}px | "
+        f"Setup: {setup_time:.2f}s, Single chunk: {capture_time:.2f}s, Total: {total_time:.2f}s")
 
     return screenshot_chunks[0]
 
@@ -184,7 +213,8 @@ class fetcher(Fetcher):
 
     async def screenshot_step(self, step_n=''):
         super().screenshot_step(step_n=step_n)
-        screenshot = await capture_full_page_async(page=self.page, screenshot_format=self.screenshot_format)
+        watch_uuid = getattr(self, 'watch_uuid', None)
+        screenshot = await capture_full_page_async(page=self.page, screenshot_format=self.screenshot_format, watch_uuid=watch_uuid, lock_viewport_elements=self.lock_viewport_elements)
 
         # Request GC immediately after screenshot to free memory
         # Screenshots can be large and browser steps take many of them
@@ -233,6 +263,7 @@ class fetcher(Fetcher):
         import playwright._impl._errors
         import time
         self.delete_browser_steps_screenshots()
+        self.watch_uuid = watch_uuid  # Store for use in screenshot_step
         response = None
 
         async with async_playwright() as p:
@@ -318,13 +349,8 @@ class fetcher(Fetcher):
                     logger.error(f"Error fetching FavIcon info {str(e)}, continuing.")
 
             if self.status_code != 200 and not ignore_status_codes:
-                screenshot = await capture_full_page_async(self.page, screenshot_format=self.screenshot_format)
-                # Cleanup before raising to prevent memory leak
-                await self.page.close()
-                await context.close()
-                await browser.close()
-                # Force garbage collection to release Playwright resources immediately
-                gc.collect()
+                screenshot = await capture_full_page_async(self.page, screenshot_format=self.screenshot_format, watch_uuid=watch_uuid, lock_viewport_elements=self.lock_viewport_elements)
+                # Finally block will handle cleanup
                 raise Non200ErrorCodeReceived(url=url, status_code=self.status_code, screenshot=screenshot)
 
             if not empty_pages_are_a_change and len((await self.page.content()).strip()) == 0:
@@ -337,7 +363,11 @@ class fetcher(Fetcher):
             try:
                 # Run Browser Steps here
                 if self.browser_steps_get_valid_steps():
-                    await self.iterate_browser_steps(start_url=url)
+                    try:
+                        await self.iterate_browser_steps(start_url=url)
+                    except BrowserStepsStepException:
+                        # Finally block will handle cleanup
+                        raise
 
                     await self.page.wait_for_timeout(extra_wait * 1000)
 
@@ -374,44 +404,51 @@ class fetcher(Fetcher):
                 # which will significantly increase the IO size between the server and client, it's recommended to use the lowest
                 # acceptable screenshot quality here
                 # The actual screenshot - this always base64 and needs decoding! horrible! huge CPU usage
-                self.screenshot = await capture_full_page_async(page=self.page, screenshot_format=self.screenshot_format)
+                self.screenshot = await capture_full_page_async(page=self.page, screenshot_format=self.screenshot_format, watch_uuid=watch_uuid, lock_viewport_elements=self.lock_viewport_elements)
+
+                # Force aggressive memory cleanup - screenshots are large and base64 decode creates temporary buffers
+                await self.page.request_gc()
+                gc.collect()
 
             except ScreenshotUnavailable:
                 # Re-raise screenshot unavailable exceptions
-                raise
-            except Exception as e:
-                # It's likely the screenshot was too long/big and something crashed
                 raise ScreenshotUnavailable(url=url, status_code=self.status_code)
+
             finally:
-                # Request garbage collection one more time before closing
+                # Clean up resources properly with timeouts to prevent hanging
                 try:
-                    await self.page.request_gc()
-                except:
-                    pass
-                
-                # Clean up resources properly
-                try:
-                    await self.page.request_gc()
-                except:
-                    pass
+                    if hasattr(self, 'page') and self.page:
+                        await self.page.request_gc()
+                        await asyncio.wait_for(self.page.close(), timeout=5.0)
+                        logger.debug(f"Successfully closed page for {url}")
+                except asyncio.TimeoutError:
+                    logger.warning(f"Timed out closing page for {url} (5s)")
+                except Exception as e:
+                    logger.warning(f"Error closing page for {url}: {e}")
+                finally:
+                    self.page = None
 
                 try:
-                    await self.page.close()
-                except:
-                    pass
-                self.page = None
+                    if context:
+                        await asyncio.wait_for(context.close(), timeout=5.0)
+                        logger.debug(f"Successfully closed context for {url}")
+                except asyncio.TimeoutError:
+                    logger.warning(f"Timed out closing context for {url} (5s)")
+                except Exception as e:
+                    logger.warning(f"Error closing context for {url}: {e}")
+                finally:
+                    context = None
 
                 try:
-                    await context.close()
-                except:
-                    pass
-                context = None
-
-                try:
-                    await browser.close()
-                except:
-                    pass
-                browser = None
+                    if browser:
+                        await asyncio.wait_for(browser.close(), timeout=5.0)
+                        logger.debug(f"Successfully closed browser connection for {url}")
+                except asyncio.TimeoutError:
+                    logger.warning(f"Timed out closing browser connection for {url} (5s)")
+                except Exception as e:
+                    logger.warning(f"Error closing browser for {url}: {e}")
+                finally:
+                    browser = None
 
                 # Force Python GC to release Playwright resources immediately
                 # Playwright objects can have circular references that delay cleanup

changedetectionio/content_fetchers/puppeteer.py

@@ -1,4 +1,5 @@
 import asyncio
+import gc
 import json
 import os
 import websockets.exceptions
@@ -20,18 +21,20 @@ from changedetectionio.content_fetchers.exceptions import PageUnloadable, Non200
 # Screenshots also travel via the ws:// (websocket) meaning that the binary data is base64 encoded
 # which will significantly increase the IO size between the server and client, it's recommended to use the lowest
 # acceptable screenshot quality here
-async def capture_full_page(page, screenshot_format='JPEG'):
+async def capture_full_page(page, screenshot_format='JPEG', watch_uuid=None, lock_viewport_elements=False):
     import os
     import time
-    import multiprocessing
 
     start = time.time()
+    watch_info = f"[{watch_uuid}] " if watch_uuid else ""
 
+    setup_start = time.time()
     page_height = await page.evaluate("document.documentElement.scrollHeight")
     page_width = await page.evaluate("document.documentElement.scrollWidth")
     original_viewport = page.viewport
+    dimensions_time = time.time() - setup_start
 
-    logger.debug(f"Puppeteer viewport size {page.viewport} page height {page_height} page width {page_width}")
+    logger.debug(f"{watch_info}Puppeteer viewport size {page.viewport} page height {page_height} page width {page_width} (got dimensions in {dimensions_time:.2f}s)")
 
     # Bug 3 in Playwright screenshot handling
     # Some bug where it gives the wrong screenshot size, but making a request with the clip set first seems to solve it
@@ -50,20 +53,35 @@ async def capture_full_page(page, screenshot_format='JPEG'):
     screenshot_chunks = []
     y = 0
     elements_locked = False
-    if page_height > page.viewport['height']:
-        # Lock all element dimensions BEFORE screenshot to prevent CSS media queries from resizing
-        # capture_full_page() changes viewport height which triggers @media (min-height) rules
+
+    # Only lock viewport elements if explicitly enabled (for image_ssim_diff processor)
+    # This prevents headers/ads from resizing when viewport changes
+    if lock_viewport_elements and page_height > page.viewport['height']:
+        lock_start = time.time()
         lock_elements_js_path = os.path.join(os.path.dirname(__file__), 'res', 'lock-elements-sizing.js')
+        file_read_start = time.time()
         with open(lock_elements_js_path, 'r') as f:
             lock_elements_js = f.read()
-        await page.evaluate(lock_elements_js)
-        elements_locked = True
-        logger.debug("Element dimensions locked before screenshot capture")
+        file_read_time = time.time() - file_read_start
 
+        evaluate_start = time.time()
+        await page.evaluate(lock_elements_js)
+        evaluate_time = time.time() - evaluate_start
+
+        elements_locked = True
+        lock_time = time.time() - lock_start
+        logger.debug(f"{watch_info}Viewport element locking enabled - File read: {file_read_time:.3f}s, Browser evaluate: {evaluate_time:.2f}s, Total: {lock_time:.2f}s")
+
+    if page_height > page.viewport['height']:
         if page_height < step_size:
             step_size = page_height # Incase page is bigger than default viewport but smaller than proposed step size
+        viewport_start = time.time()
         await page.setViewport({'width': page.viewport['width'], 'height': step_size})
+        viewport_time = time.time() - viewport_start
+        logger.debug(f"{watch_info}Viewport changed to {page.viewport['width']}x{step_size} (took {viewport_time:.2f}s)")
 
+    capture_start = time.time()
+    chunk_times = []
     while y < min(page_height, SCREENSHOT_MAX_TOTAL_HEIGHT):
         # better than scrollTo incase they override it in the page
         await page.evaluate(
@@ -82,7 +100,11 @@ async def capture_full_page(page, screenshot_format='JPEG'):
         if screenshot_type == 'jpeg':
             screenshot_kwargs['quality'] = screenshot_quality
 
+        chunk_start = time.time()
         screenshot_chunks.append(await page.screenshot(**screenshot_kwargs))
+        chunk_time = time.time() - chunk_start
+        chunk_times.append(chunk_time)
+        logger.debug(f"{watch_info}Chunk {len(screenshot_chunks)} captured in {chunk_time:.2f}s")
         y += step_size
 
     await page.setViewport({'width': original_viewport['width'], 'height': original_viewport['height']})
@@ -93,26 +115,53 @@ async def capture_full_page(page, screenshot_format='JPEG'):
         with open(unlock_elements_js_path, 'r') as f:
             unlock_elements_js = f.read()
         await page.evaluate(unlock_elements_js)
-        logger.debug("Element dimensions unlocked after screenshot capture")
+        logger.debug(f"{watch_info}Element dimensions unlocked after screenshot capture")
+
+    capture_time = time.time() - capture_start
+    total_capture_time = sum(chunk_times)
+    logger.debug(f"{watch_info}All {len(screenshot_chunks)} chunks captured in {capture_time:.2f}s (total chunk time: {total_capture_time:.2f}s)")
 
     if len(screenshot_chunks) > 1:
-        # Always use spawn for thread safety - consistent behavior in tests and production
-        from changedetectionio.content_fetchers.screenshot_handler import stitch_images_worker
-        logger.debug(f"Screenshot stitching {len(screenshot_chunks)} chunks together")
+        stitch_start = time.time()
+        logger.debug(f"{watch_info}Starting stitching of {len(screenshot_chunks)} chunks")
+
+        # Always use spawn subprocess for ANY stitching (2+ chunks)
+        # PIL allocates at C level and Python GC never releases it - subprocess exit forces OS to reclaim
+        # Trade-off: 35MB resource_tracker vs 500MB+ PIL leak in main process
+        from changedetectionio.content_fetchers.screenshot_handler import stitch_images_worker_raw_bytes
+        import multiprocessing
+        import struct
+
         ctx = multiprocessing.get_context('spawn')
         parent_conn, child_conn = ctx.Pipe()
-        p = ctx.Process(target=stitch_images_worker, args=(child_conn, screenshot_chunks, page_height, SCREENSHOT_MAX_TOTAL_HEIGHT))
+        p = ctx.Process(target=stitch_images_worker_raw_bytes, args=(child_conn, page_height, SCREENSHOT_MAX_TOTAL_HEIGHT))
         p.start()
+
+        # Send via raw bytes (no pickle)
+        parent_conn.send_bytes(struct.pack('I', len(screenshot_chunks)))
+        for chunk in screenshot_chunks:
+            parent_conn.send_bytes(chunk)
+
         screenshot = parent_conn.recv_bytes()
         p.join()
-        logger.debug(
-            f"Screenshot (chunked/stitched) - Page height: {page_height} Capture height: {SCREENSHOT_MAX_TOTAL_HEIGHT} - Stitched together in {time.time() - start:.2f}s")
 
-        screenshot_chunks = None
+        parent_conn.close()
+        child_conn.close()
+        del p, parent_conn, child_conn
+
+        stitch_time = time.time() - stitch_start
+        total_time = time.time() - start
+        setup_time = total_time - capture_time - stitch_time
+        logger.debug(
+            f"{watch_info}Screenshot complete - Page height: {page_height}px, Capture height: {SCREENSHOT_MAX_TOTAL_HEIGHT}px | "
+            f"Setup: {setup_time:.2f}s, Capture: {capture_time:.2f}s, Stitching: {stitch_time:.2f}s, Total: {total_time:.2f}s")
         return screenshot
 
+    total_time = time.time() - start
+    setup_time = total_time - capture_time
     logger.debug(
-        f"Screenshot Page height: {page_height} Capture height: {SCREENSHOT_MAX_TOTAL_HEIGHT} - Stitched together in {time.time() - start:.2f}s")
+        f"{watch_info}Screenshot complete - Page height: {page_height}px, Capture height: {SCREENSHOT_MAX_TOTAL_HEIGHT}px | "
+        f"Setup: {setup_time:.2f}s, Single chunk: {capture_time:.2f}s, Total: {total_time:.2f}s")
     return screenshot_chunks[0]
 
 
@@ -173,19 +222,36 @@ class fetcher(Fetcher):
                 self.browser_connection_url += f"{r}--proxy-server={proxy_url}"
 
     async def quit(self, watch=None):
-        try:
-            await self.page.close()
-            del self.page
-        except Exception as e:
-            pass
+        watch_uuid = watch.get('uuid') if watch else 'unknown'
 
+        # Close page
         try:
-            await self.browser.close()
-            del self.browser
+            if hasattr(self, 'page') and self.page:
+                await asyncio.wait_for(self.page.close(), timeout=5.0)
+                logger.debug(f"[{watch_uuid}] Page closed successfully")
+        except asyncio.TimeoutError:
+            logger.warning(f"[{watch_uuid}] Timed out closing page (5s)")
         except Exception as e:
-            pass
+            logger.warning(f"[{watch_uuid}] Error closing page: {e}")
+        finally:
+            self.page = None
 
-        logger.info("Cleanup puppeteer complete.")
+        # Close browser connection
+        try:
+            if hasattr(self, 'browser') and self.browser:
+                await asyncio.wait_for(self.browser.close(), timeout=5.0)
+                logger.debug(f"[{watch_uuid}] Browser closed successfully")
+        except asyncio.TimeoutError:
+            logger.warning(f"[{watch_uuid}] Timed out closing browser (5s)")
+        except Exception as e:
+            logger.warning(f"[{watch_uuid}] Error closing browser: {e}")
+        finally:
+            self.browser = None
+
+        logger.info(f"[{watch_uuid}] Cleanup puppeteer complete")
+
+        # Force garbage collection to release resources
+        gc.collect()
 
     async def fetch_page(self,
                          current_include_filters,
@@ -215,9 +281,11 @@ class fetcher(Fetcher):
         # Connect directly using the specified browser_ws_endpoint
         # @todo timeout
         try:
+            logger.debug(f"[{watch_uuid}] Connecting to browser at {self.browser_connection_url}")
             self.browser = await pyppeteer_instance.connect(browserWSEndpoint=self.browser_connection_url,
                                                             ignoreHTTPSErrors=True
                                                             )
+            logger.debug(f"[{watch_uuid}] Browser connected successfully")
         except websockets.exceptions.InvalidStatusCode as e:
             raise BrowserConnectError(msg=f"Error while trying to connect the browser, Code {e.status_code} (check your access, whitelist IP, password etc)")
         except websockets.exceptions.InvalidURI:
@@ -226,7 +294,18 @@ class fetcher(Fetcher):
             raise BrowserConnectError(msg=f"Error connecting to the browser - Exception '{str(e)}'")
 
         # more reliable is to just request a new page
-        self.page = await self.browser.newPage()
+        try:
+            logger.debug(f"[{watch_uuid}] Creating new page")
+            self.page = await self.browser.newPage()
+            logger.debug(f"[{watch_uuid}] Page created successfully")
+        except Exception as e:
+            logger.error(f"[{watch_uuid}] Failed to create new page: {e}")
+            # Browser is connected but page creation failed - must cleanup browser
+            try:
+                await asyncio.wait_for(self.browser.close(), timeout=3.0)
+            except Exception as cleanup_error:
+                logger.error(f"[{watch_uuid}] Failed to cleanup browser after page creation failure: {cleanup_error}")
+            raise
         
         # Add console handler to capture console.log from favicon fetcher
         #self.page.on('console', lambda msg: logger.debug(f"Browser console [{msg.type}]: {msg.text}"))
@@ -295,6 +374,12 @@ class fetcher(Fetcher):
             w = extra_wait - 2 if extra_wait > 4 else 2
             logger.debug(f"Waiting {w} seconds before calling Page.stopLoading...")
             await asyncio.sleep(w)
+
+            # Check if page still exists (might have been closed due to error during sleep)
+            if not self.page or not hasattr(self.page, '_client'):
+                logger.debug("Page already closed, skipping stopLoading")
+                return
+
             logger.debug("Issuing stopLoading command...")
             await self.page._client.send('Page.stopLoading')
             logger.debug("stopLoading command sent!")
@@ -320,7 +405,9 @@ class fetcher(Fetcher):
             asyncio.create_task(handle_frame_navigation())
             response = await self.page.goto(url, timeout=0)
             await asyncio.sleep(1 + extra_wait)
-            await self.page._client.send('Page.stopLoading')
+            # Check if page still exists before sending command
+            if self.page and hasattr(self.page, '_client'):
+                await self.page._client.send('Page.stopLoading')
 
             if response:
                 break
@@ -357,7 +444,7 @@ class fetcher(Fetcher):
                 logger.error(f"Error fetching FavIcon info {str(e)}, continuing.")
 
         if self.status_code != 200 and not ignore_status_codes:
-            screenshot = await capture_full_page(page=self.page, screenshot_format=self.screenshot_format)
+            screenshot = await capture_full_page(page=self.page, screenshot_format=self.screenshot_format, watch_uuid=watch_uuid, lock_viewport_elements=self.lock_viewport_elements)
 
             raise Non200ErrorCodeReceived(url=url, status_code=self.status_code, screenshot=screenshot)
 
@@ -387,7 +474,11 @@ class fetcher(Fetcher):
 
         # Now take screenshot (scrolling may trigger layout changes, but measurements are already captured)
         logger.debug(f"Screenshot format {self.screenshot_format}")
-        self.screenshot = await capture_full_page(page=self.page, screenshot_format=self.screenshot_format)
+        self.screenshot = await capture_full_page(page=self.page, screenshot_format=self.screenshot_format, watch_uuid=watch_uuid, lock_viewport_elements=self.lock_viewport_elements)
+
+        # Force garbage collection - pyppeteer base64 decode creates temporary buffers
+        import gc
+        gc.collect()
         self.xpath_data = await self.page.evaluate(XPATH_ELEMENT_JS, {
             "visualselector_xpath_selectors": visualselector_xpath_selectors,
             "max_height": MAX_TOTAL_HEIGHT

changedetectionio/content_fetchers/requests.py

@@ -55,6 +55,26 @@ class fetcher(Fetcher):
 
         session = requests.Session()
 
+        # Configure retry adapter for low-level network errors only
+        # Retries connection timeouts, read timeouts, connection resets - not HTTP status codes
+        # Especially helpful in parallel test execution when servers are slow/overloaded
+        # Configurable via REQUESTS_RETRY_MAX_COUNT (default: 3 attempts)
+        from requests.adapters import HTTPAdapter
+        from urllib3.util.retry import Retry
+
+        max_retries = int(os.getenv("REQUESTS_RETRY_MAX_COUNT", "6"))
+        retry_strategy = Retry(
+            total=max_retries,
+            connect=max_retries,  # Retry connection timeouts
+            read=max_retries,     # Retry read timeouts
+            status=0,             # Don't retry on HTTP status codes
+            backoff_factor=0.5,   # Wait 0.3s, 0.6s, 1.2s between retries
+            allowed_methods=["HEAD", "GET", "OPTIONS", "POST"],
+            raise_on_status=False
+        )
+        adapter = HTTPAdapter(max_retries=retry_strategy)
+        session.mount("http://", adapter)
+        session.mount("https://", adapter)
 
         if strtobool(os.getenv('ALLOW_FILE_URI', 'false')) and url.startswith('file://'):
             from requests_file import FileAdapter
@@ -142,10 +162,11 @@ class fetcher(Fetcher):
                   watch_uuid=None,
                   ):
         """Async wrapper that runs the synchronous requests code in a thread pool"""
-        
+
         loop = asyncio.get_event_loop()
-        
+
         # Run the synchronous _run_sync in a thread pool to avoid blocking the event loop
+        # Retry logic is handled by requests' HTTPAdapter (see _run_sync for configuration)
         await loop.run_in_executor(
             None,  # Use default ThreadPoolExecutor
             lambda: self._run_sync(

changedetectionio/content_fetchers/res/lock-elements-sizing.js

@@ -1,5 +1,5 @@
 /**
- * Lock Element Dimensions for Screenshot Capture
+ * Lock Element Dimensions for Screenshot Capture (First Viewport Only)
  *
  * THE PROBLEM:
  * When taking full-page screenshots of tall pages, Chrome/Puppeteer/Playwright need to:
@@ -10,40 +10,31 @@
  * However, changing the viewport height triggers CSS media queries like:
  *   @media (min-height: 860px) { .ad { height: 250px; } }
  *
- * This causes elements (especially ads) to resize during screenshot capture, creating a mismatch:
- * - Screenshot shows element at NEW size (after media query triggered)
- * - xpath element coordinates measured at OLD size (before viewport change)
- * - Visual selector overlays don't align with screenshot
- *
- * EXAMPLE BUG:
- * - Initial viewport: 1280x800, ad height: 138px, article position: 279px ✓
- * - Viewport changes to 1280x3809 for screenshot
- * - Media query triggers: ad expands to 250px
- * - All content below shifts down by 112px (250-138)
- * - Article now at position: 391px (279+112)
- * - But xpath data says 279px → 112px mismatch! ✗
+ * This causes elements (especially ads/headers) to resize during screenshot capture.
  *
  * THE SOLUTION:
- * Before changing viewport, lock ALL element dimensions with !important inline styles.
- * Inline styles with !important override media query CSS, preventing layout changes.
+ * Lock element dimensions in the FIRST VIEWPORT ONLY with !important inline styles.
+ * This prevents headers, navigation, and top ads from resizing when viewport changes.
+ * We only lock the visible portion because:
+ * - Most layout shifts happen in headers/navbars/top ads
+ * - Locking only visible elements is 100x+ faster (100-200 elements vs 10,000+)
+ * - Below-fold content shifts don't affect visual comparison accuracy
  *
  * WHAT THIS SCRIPT DOES:
- * 1. Iterates through every element on the page
- * 2. Captures current computed dimensions (width, height)
- * 3. Sets inline styles with !important to freeze those dimensions
+ * 1. Gets current viewport height
+ * 2. Finds elements within first viewport (top of page to bottom of screen)
+ * 3. Locks their dimensions with !important inline styles
  * 4. Disables ResizeObserver API (for JS-based resizing)
- * 5. When viewport changes for screenshot, media queries can't resize anything
- * 6. Layout remains consistent → xpath coordinates match screenshot ✓
  *
  * USAGE:
  * Execute this script BEFORE calling capture_full_page() / screenshot functions.
- * The page must be fully loaded and settled at its initial viewport size.
- * No need to restore state afterward - page is closed after screenshot.
+ * Only enabled for image_ssim_diff processor (visual comparison).
+ * Default: OFF for performance.
  *
  * PERFORMANCE:
- * - Iterates all DOM elements (can be 1000s on complex pages)
- * - Typically completes in 50-200ms
- * - One-time cost before screenshot, well worth it for coordinate accuracy
+ * - Only processes 100-300 elements (first viewport) vs 10,000+ (entire page)
+ * - Typically completes in 10-50ms
+ * - 100x+ faster than locking entire page
  *
  * @see https://github.com/dgtlmoon/changedetection.io/issues/XXXX
  */
@@ -52,11 +43,34 @@
     // Store original styles in a global WeakMap for later restoration
     window.__elementSizingRestore = new WeakMap();
 
-    // Lock ALL element dimensions to prevent media query layout changes
-    document.querySelectorAll('*').forEach(el => {
-        const computed = window.getComputedStyle(el);
-        const rect = el.getBoundingClientRect();
+    const start = performance.now();
 
+    // Get current viewport height (visible portion of page)
+    const viewportHeight = window.innerHeight;
+
+    // Get all elements and filter to FIRST VIEWPORT ONLY
+    // This dramatically reduces elements to process (100-300 vs 10,000+)
+    const allElements = Array.from(document.querySelectorAll('*'));
+
+    // BATCH READ PHASE: Get bounding rects and filter to viewport
+    const measurements = allElements.map(el => {
+        const rect = el.getBoundingClientRect();
+        const computed = window.getComputedStyle(el);
+
+        // Only lock elements in the first viewport (visible on initial page load)
+        // rect.top < viewportHeight means element starts within visible area
+        const inViewport = rect.top < viewportHeight && rect.top >= 0;
+        const hasSize = rect.height > 0 && rect.width > 0;
+
+        return inViewport && hasSize ? { el, computed, rect } : null;
+    }).filter(Boolean);  // Remove null entries
+
+    const elapsed = performance.now() - start;
+    console.log(`Locked first viewport elements: ${measurements.length} of ${allElements.length} total elements (viewport height: ${viewportHeight}px, took ${elapsed.toFixed(0)}ms)`);
+
+    // BATCH WRITE PHASE: Apply all inline styles without triggering layout
+    // No interleaved reads means browser can optimize style application
+    measurements.forEach(({el, computed, rect}) => {
         // Save original inline style values BEFORE locking
         const properties = ['height', 'min-height', 'max-height', 'width', 'min-width', 'max-width'];
         const originalStyles = {};
@@ -89,5 +103,5 @@
         disconnect() {}
     };
 
-    console.log('✓ Element dimensions locked to prevent media query changes during screenshot');
+    console.log(`✓ Element dimensions locked (${measurements.length} elements) to prevent media query changes during screenshot`);
 })();

changedetectionio/content_fetchers/screenshot_handler.py

@@ -8,92 +8,42 @@ from loguru import logger
 
 from changedetectionio.content_fetchers import SCREENSHOT_MAX_HEIGHT_DEFAULT, SCREENSHOT_DEFAULT_QUALITY
 
-# Cache font to avoid loading on every stitch
-_cached_font = None
-
-def _get_caption_font():
-    """Get or create cached font for caption text."""
-    global _cached_font
-    if _cached_font is None:
-        from PIL import ImageFont
-        try:
-            _cached_font = ImageFont.truetype("arial.ttf", 35)
-        except IOError:
-            _cached_font = ImageFont.load_default()
-    return _cached_font
-
-
-def stitch_images_inline(chunks_bytes, original_page_height, capture_height):
-    """
-    Stitch image chunks together inline (no multiprocessing).
-    Optimized for small number of chunks (2-3) to avoid process creation overhead.
-
-    Args:
-        chunks_bytes: List of JPEG image bytes
-        original_page_height: Original page height in pixels
-        capture_height: Maximum capture height
-
-    Returns:
-        bytes: Stitched JPEG image
-    """
-    import os
-    import io
-    from PIL import Image, ImageDraw
-
-    # Load images from byte chunks
-    images = [Image.open(io.BytesIO(b)) for b in chunks_bytes]
-    total_height = sum(im.height for im in images)
-    max_width = max(im.width for im in images)
-
-    # Create stitched image
-    stitched = Image.new('RGB', (max_width, total_height))
-    y_offset = 0
-    for im in images:
-        stitched.paste(im, (0, y_offset))
-        y_offset += im.height
-        im.close()  # Close immediately after pasting
-
-    # Draw caption only if page was trimmed
-    if original_page_height > capture_height:
-        draw = ImageDraw.Draw(stitched)
-        caption_text = f"WARNING: Screenshot was {original_page_height}px but trimmed to {capture_height}px because it was too long"
-        padding = 10
-        font = _get_caption_font()
-
-        bbox = draw.textbbox((0, 0), caption_text, font=font)
-        text_width = bbox[2] - bbox[0]
-        text_height = bbox[3] - bbox[1]
-
-        # Draw white background rectangle
-        draw.rectangle([(0, 0), (max_width, text_height + 2 * padding)], fill=(255, 255, 255))
-
-        # Draw text centered
-        text_x = (max_width - text_width) // 2
-        draw.text((text_x, padding), caption_text, font=font, fill=(255, 0, 0))
-
-    # Encode to JPEG
-    output = io.BytesIO()
-    stitched.save(output, format="JPEG", quality=int(os.getenv("SCREENSHOT_QUALITY", SCREENSHOT_DEFAULT_QUALITY)), optimize=True)
-    result = output.getvalue()
-
-    # Cleanup
-    stitched.close()
-
-    return result
-
-
-def stitch_images_worker(pipe_conn, chunks_bytes, original_page_height, capture_height):
+def stitch_images_worker_raw_bytes(pipe_conn, original_page_height, capture_height):
     """
     Stitch image chunks together in a separate process.
-    Used for large number of chunks (4+) to avoid blocking the main event loop.
+
+    Uses spawn multiprocessing to isolate PIL's C-level memory allocation.
+    When the subprocess exits, the OS reclaims ALL memory including C-level allocations
+    that Python's GC cannot release. This prevents the ~50MB per stitch from accumulating
+    in the main process.
+
+    Trade-off: Adds 35MB resource_tracker subprocess, but prevents 500MB+ memory leak
+    in main process (much better at scale: 35GB vs 500GB for 1000 instances).
+
+    Args:
+        pipe_conn: Pipe connection to receive data and send result
+        original_page_height: Original page height in pixels
+        capture_height: Maximum capture height
     """
     import os
     import io
+    import struct
     from PIL import Image, ImageDraw, ImageFont
 
     try:
+        # Receive chunk count as 4-byte integer (no pickle!)
+        count_bytes = pipe_conn.recv_bytes()
+        chunk_count = struct.unpack('I', count_bytes)[0]
+
+        # Receive each chunk as raw bytes (no pickle!)
+        chunks_bytes = []
+        for _ in range(chunk_count):
+            chunks_bytes.append(pipe_conn.recv_bytes())
+
         # Load images from byte chunks
         images = [Image.open(io.BytesIO(b)) for b in chunks_bytes]
+        del chunks_bytes
+
         total_height = sum(im.height for im in images)
         max_width = max(im.width for im in images)
 
@@ -103,15 +53,14 @@ def stitch_images_worker(pipe_conn, chunks_bytes, original_page_height, capture_
         for im in images:
             stitched.paste(im, (0, y_offset))
             y_offset += im.height
-            im.close()  # Close immediately after pasting
+            im.close()
+        del images
 
         # Draw caption only if page was trimmed
         if original_page_height > capture_height:
             draw = ImageDraw.Draw(stitched)
             caption_text = f"WARNING: Screenshot was {original_page_height}px but trimmed to {capture_height}px because it was too long"
             padding = 10
-
-            # Try to load font
             try:
                 font = ImageFont.truetype("arial.ttf", 35)
             except IOError:
@@ -120,23 +69,26 @@ def stitch_images_worker(pipe_conn, chunks_bytes, original_page_height, capture_
             bbox = draw.textbbox((0, 0), caption_text, font=font)
             text_width = bbox[2] - bbox[0]
             text_height = bbox[3] - bbox[1]
-
-            # Draw white background rectangle
             draw.rectangle([(0, 0), (max_width, text_height + 2 * padding)], fill=(255, 255, 255))
-
-            # Draw text centered
             text_x = (max_width - text_width) // 2
             draw.text((text_x, padding), caption_text, font=font, fill=(255, 0, 0))
 
-        # Encode and send image with optimization
+        # Encode and send
         output = io.BytesIO()
         stitched.save(output, format="JPEG", quality=int(os.getenv("SCREENSHOT_QUALITY", SCREENSHOT_DEFAULT_QUALITY)), optimize=True)
-        pipe_conn.send_bytes(output.getvalue())
+        result_bytes = output.getvalue()
 
         stitched.close()
+        del stitched
+        output.close()
+        del output
+
+        pipe_conn.send_bytes(result_bytes)
+        del result_bytes
+
     except Exception as e:
-        pipe_conn.send(f"error:{e}")
+        logger.error(f"Error in stitch_images_worker_raw_bytes: {e}")
+        error_msg = f"error:{e}".encode('utf-8')
+        pipe_conn.send_bytes(error_msg)
     finally:
         pipe_conn.close()
-
-

changedetectionio/flask_app.py

@@ -14,7 +14,7 @@ from pathlib import Path
 from changedetectionio.strtobool import strtobool
 from threading import Event
 from changedetectionio.queue_handlers import RecheckPriorityQueue, NotificationQueue
-from changedetectionio import worker_handler
+from changedetectionio import worker_pool
 
 from flask import (
     Flask,
@@ -94,6 +94,14 @@ if os.getenv('FLASK_SERVER_NAME'):
 app.config['BABEL_TRANSLATION_DIRECTORIES'] = str(Path(__file__).parent / 'translations')
 app.config['BABEL_DEFAULT_LOCALE'] = 'en_GB'
 
+# Session configuration
+# NOTE: Flask session (for locale, etc.) is separate from Flask-Login's remember-me cookie
+# - Flask session stores data like session['locale'] in a signed cookie
+# - Flask-Login's remember=True creates a separate authentication cookie
+# - Setting PERMANENT_SESSION_LIFETIME controls how long the Flask session cookie lasts
+from datetime import timedelta
+app.config['PERMANENT_SESSION_LIFETIME'] = timedelta(days=3650)  # ~10 years (effectively unlimited)
+
 #app.config["EXPLAIN_TEMPLATE_LOADING"] = True
 
 
@@ -187,7 +195,7 @@ def _jinja2_filter_format_number_locale(value: float) -> str:
 
 @app.template_global('is_checking_now')
 def _watch_is_checking_now(watch_obj, format="%Y-%m-%d %H:%M:%S"):
-    return worker_handler.is_watch_running(watch_obj['uuid'])
+    return worker_pool.is_watch_running(watch_obj['uuid'])
 
 @app.template_global('get_watch_queue_position')
 def _get_watch_queue_position(watch_obj):
@@ -198,13 +206,13 @@ def _get_watch_queue_position(watch_obj):
 @app.template_global('get_current_worker_count')
 def _get_current_worker_count():
     """Get the current number of operational workers"""
-    return worker_handler.get_worker_count()
+    return worker_pool.get_worker_count()
 
 @app.template_global('get_worker_status_info')
 def _get_worker_status_info():
     """Get detailed worker status information for display"""
-    status = worker_handler.get_worker_status()
-    running_uuids = worker_handler.get_running_uuids()
+    status = worker_pool.get_worker_status()
+    running_uuids = worker_pool.get_running_uuids()
     
     return {
         'count': status['worker_count'],
@@ -258,6 +266,47 @@ def _jinja2_filter_seconds_precise(timestamp):
 
     return format(int(time.time()-timestamp), ',d')
 
+@app.template_filter('format_duration')
+def _jinja2_filter_format_duration(seconds):
+    """Format a duration in seconds into human readable string like '5 days, 3 hours, 30 minutes'"""
+    from datetime import timedelta
+
+    if not seconds or seconds < 0:
+        return gettext('0 seconds')
+
+    td = timedelta(seconds=int(seconds))
+
+    # Calculate components
+    years = td.days // 365
+    remaining_days = td.days % 365
+    months = remaining_days // 30
+    remaining_days = remaining_days % 30
+    weeks = remaining_days // 7
+    days = remaining_days % 7
+
+    hours = td.seconds // 3600
+    minutes = (td.seconds % 3600) // 60
+    secs = td.seconds % 60
+
+    # Build parts list
+    parts = []
+    if years > 0:
+        parts.append(f"{years} {gettext('year') if years == 1 else gettext('years')}")
+    if months > 0:
+        parts.append(f"{months} {gettext('month') if months == 1 else gettext('months')}")
+    if weeks > 0:
+        parts.append(f"{weeks} {gettext('week') if weeks == 1 else gettext('weeks')}")
+    if days > 0:
+        parts.append(f"{days} {gettext('day') if days == 1 else gettext('days')}")
+    if hours > 0:
+        parts.append(f"{hours} {gettext('hour') if hours == 1 else gettext('hours')}")
+    if minutes > 0:
+        parts.append(f"{minutes} {gettext('minute') if minutes == 1 else gettext('minutes')}")
+    if secs > 0 or not parts:
+        parts.append(f"{secs} {gettext('second') if secs == 1 else gettext('seconds')}")
+
+    return ", ".join(parts)
+
 @app.template_filter('fetcher_status_icons')
 def _jinja2_filter_fetcher_status_icons(fetcher_name):
     """Get status icon HTML for a given fetcher.
@@ -393,7 +442,10 @@ def changedetection_app(config=None, datastore_o=None):
     # so far just for read-only via tests, but this will be moved eventually to be the main source
     # (instead of the global var)
     app.config['DATASTORE'] = datastore_o
-    
+
+    # Store batch mode flag to skip background threads when running in batch mode
+    app.config['batch_mode'] = config.get('batch_mode', False) if config else False
+
     # Store the signal in the app config to ensure it's accessible everywhere
     app.config['watch_check_update_SIGNAL'] = watch_check_update
 
@@ -550,6 +602,9 @@ def changedetection_app(config=None, datastore_o=None):
 
         # Validate the locale against available languages
         if locale in language_codes:
+            # Make session permanent so language preference persists across browser sessions
+            # NOTE: This is the Flask session cookie (separate from Flask-Login's remember-me auth cookie)
+            session.permanent = True
             session['locale'] = locale
 
             # CRITICAL: Flask-Babel caches the locale in the request context (ctx.babel_locale)
@@ -689,10 +744,10 @@ def changedetection_app(config=None, datastore_o=None):
             favicon_filename = watch.get_favicon_filename()
             if favicon_filename:
                 # Use cached MIME type detection
-                filepath = os.path.join(watch.watch_data_dir, favicon_filename)
+                filepath = os.path.join(watch.data_dir, favicon_filename)
                 mime = get_favicon_mime_type(filepath)
 
-                response = make_response(send_from_directory(watch.watch_data_dir, favicon_filename))
+                response = make_response(send_from_directory(watch.data_dir, favicon_filename))
                 response.headers['Content-type'] = mime
                 response.headers['Cache-Control'] = 'max-age=300, must-revalidate'  # Cache for 5 minutes, then revalidate
                 return response
@@ -787,13 +842,15 @@ def changedetection_app(config=None, datastore_o=None):
 
     # watchlist UI buttons etc
     import changedetectionio.blueprint.ui as ui
-    app.register_blueprint(ui.construct_blueprint(datastore, update_q, worker_handler, queuedWatchMetaData, watch_check_update))
+    app.register_blueprint(ui.construct_blueprint(datastore, update_q, worker_pool, queuedWatchMetaData, watch_check_update))
 
     import changedetectionio.blueprint.watchlist as watchlist
     app.register_blueprint(watchlist.construct_blueprint(datastore=datastore, update_q=update_q, queuedWatchMetaData=queuedWatchMetaData), url_prefix='')
 
     # Initialize Socket.IO server conditionally based on settings
-    socket_io_enabled = datastore.data['settings']['application']['ui'].get('socket_io_enabled', True)
+    socket_io_enabled = datastore.data['settings']['application'].get('ui', {}).get('socket_io_enabled', True)
+    if socket_io_enabled and app.config.get('batch_mode'):
+        socket_io_enabled = False
     if socket_io_enabled:
         from changedetectionio.realtime.socket_server import init_socketio
         global socketio_server
@@ -822,10 +879,10 @@ def changedetection_app(config=None, datastore_o=None):
         expected_workers = int(os.getenv("FETCH_WORKERS", datastore.data['settings']['requests']['workers']))
         
         # Get basic status
-        status = worker_handler.get_worker_status()
+        status = worker_pool.get_worker_status()
         
         # Perform health check
-        health_result = worker_handler.check_worker_health(
+        health_result = worker_pool.check_worker_health(
             expected_count=expected_workers,
             update_q=update_q,
             notification_q=notification_q,
@@ -889,16 +946,31 @@ def changedetection_app(config=None, datastore_o=None):
     # Can be overridden by ENV or use the default settings
     n_workers = int(os.getenv("FETCH_WORKERS", datastore.data['settings']['requests']['workers']))
     logger.info(f"Starting {n_workers} workers during app initialization")
-    worker_handler.start_workers(n_workers, update_q, notification_q, app, datastore)
+    worker_pool.start_workers(n_workers, update_q, notification_q, app, datastore)
 
-    # @todo handle ctrl break
-    ticker_thread = threading.Thread(target=ticker_thread_check_time_launch_checks, daemon=True, name="TickerThread-ScheduleChecker").start()
-    threading.Thread(target=notification_runner, daemon=True, name="NotificationRunner").start()
+    # Skip background threads in batch mode (just process queue and exit)
+    batch_mode = app.config.get('batch_mode', False)
+    if not batch_mode:
+        # @todo handle ctrl break
+        ticker_thread = threading.Thread(target=ticker_thread_check_time_launch_checks, daemon=True, name="TickerThread-ScheduleChecker").start()
 
-    in_pytest = "pytest" in sys.modules or "PYTEST_CURRENT_TEST" in os.environ
-    # Check for new release version, but not when running in test/build or pytest
-    if not os.getenv("GITHUB_REF", False) and not strtobool(os.getenv('DISABLE_VERSION_CHECK', 'no')) and not in_pytest:
-        threading.Thread(target=check_for_new_version, daemon=True, name="VersionChecker").start()
+        # Start configurable number of notification workers (default 1)
+        notification_workers = int(os.getenv("NOTIFICATION_WORKERS", "1"))
+        for i in range(notification_workers):
+            threading.Thread(
+                target=notification_runner,
+                args=(i,),
+                daemon=True,
+                name=f"NotificationRunner-{i}"
+            ).start()
+        logger.info(f"Started {notification_workers} notification worker(s)")
+
+        in_pytest = "pytest" in sys.modules or "PYTEST_CURRENT_TEST" in os.environ
+        # Check for new release version, but not when running in test/build or pytest
+        if not os.getenv("GITHUB_REF", False) and not strtobool(os.getenv('DISABLE_VERSION_CHECK', 'no')) and not in_pytest:
+            threading.Thread(target=check_for_new_version, daemon=True, name="VersionChecker").start()
+    else:
+        logger.info("Batch mode: Skipping ticker thread, notification runner, and version checker")
 
     # Return the Flask app - the Socket.IO will be attached to it but initialized separately
     # This avoids circular dependencies
@@ -933,14 +1005,14 @@ def check_for_new_version():
         app.config.exit.wait(86400)
 
 
-def notification_runner():
+def notification_runner(worker_id=0):
     global notification_debug_log
     from datetime import datetime
     import json
     with app.app_context():
         while not app.config.exit.is_set():
             try:
-                # At the moment only one thread runs (single runner)
+                # Multiple workers can run concurrently (configurable via NOTIFICATION_WORKERS)
                 n_object = notification_q.get(block=False)
             except queue.Empty:
                 app.config.exit.wait(1)
@@ -966,7 +1038,7 @@ def notification_runner():
                         sent_obj = process_notification(n_object, datastore)
 
                 except Exception as e:
-                    logger.error(f"Watch URL: {n_object['watch_url']}  Error {str(e)}")
+                    logger.error(f"Notification worker {worker_id} - Watch URL: {n_object['watch_url']}  Error {str(e)}")
 
                     # UUID wont be present when we submit a 'test' from the global settings
                     if 'uuid' in n_object:
@@ -1007,7 +1079,7 @@ def ticker_thread_check_time_launch_checks():
         now = time.time()
         if now - last_health_check > 60:
             expected_workers = int(os.getenv("FETCH_WORKERS", datastore.data['settings']['requests']['workers']))
-            health_result = worker_handler.check_worker_health(
+            health_result = worker_pool.check_worker_health(
                 expected_count=expected_workers,
                 update_q=update_q,
                 notification_q=notification_q,
@@ -1026,7 +1098,7 @@ def ticker_thread_check_time_launch_checks():
             continue
 
         # Get a list of watches by UUID that are currently fetching data
-        running_uuids = worker_handler.get_running_uuids()
+        running_uuids = worker_pool.get_running_uuids()
 
         # Build set of queued UUIDs once for O(1) lookup instead of O(n) per watch
         queued_uuids = {q_item.item['uuid'] for q_item in update_q.queue}
@@ -1132,7 +1204,7 @@ def ticker_thread_check_time_launch_checks():
                     priority = int(time.time())
 
                     # Into the queue with you
-                    queued_successfully = worker_handler.queue_item_async_safe(update_q,
+                    queued_successfully = worker_pool.queue_item_async_safe(update_q,
                                                                                queuedWatchMetaData.PrioritizedItem(priority=priority,
                                                                                                                    item={'uuid': uuid})
                                                                                )

changedetectionio/forms.py

@@ -730,7 +730,7 @@ class quickWatchForm(Form):
     url = fields.URLField(_l('URL'), validators=[validateURL()])
     tags = StringTagUUID(_l('Group tag'), validators=[validators.Optional()])
     watch_submit_button = SubmitField(_l('Watch'), render_kw={"class": "pure-button pure-button-primary"})
-    processor = RadioField(_l('Processor'), choices=lambda: processors.available_processors(), default="text_json_diff")
+    processor = RadioField(_l('Processor'), choices=lambda: processors.available_processors(), default=processors.get_default_processor)
     edit_and_watch_submit_button = SubmitField(_l('Edit > Watch'), render_kw={"class": "pure-button pure-button-primary"})
 
 
@@ -749,7 +749,7 @@ class commonSettingsForm(Form):
     notification_format = SelectField(_l('Notification format'), choices=list(valid_notification_formats.items()))
     notification_title = StringField(_l('Notification Title'), default='ChangeDetection.io Notification - {{ watch_url }}', validators=[validators.Optional(), ValidateJinja2Template()])
     notification_urls = StringListField(_l('Notification URL List'), validators=[validators.Optional(), ValidateAppRiseServers(), ValidateJinja2Template()])
-    processor = RadioField( label=_l("Processor - What do you want to achieve?"), choices=lambda: processors.available_processors(), default="text_json_diff")
+    processor = RadioField( label=_l("Processor - What do you want to achieve?"), choices=lambda: processors.available_processors(), default=processors.get_default_processor)
     scheduler_timezone_default = StringField(_l("Default timezone for watch check scheduler"), render_kw={"list": "timezones"}, validators=[validateTimeZoneName()])
     webdriver_delay = IntegerField(_l('Wait seconds before extracting text'), validators=[validators.Optional(), validators.NumberRange(min=1, message=_l("Should contain one or more seconds"))])
 
@@ -763,7 +763,7 @@ class commonSettingsForm(Form):
 
 
 class importForm(Form):
-    processor = RadioField(_l('Processor'), choices=lambda: processors.available_processors(), default="text_json_diff")
+    processor = RadioField(_l('Processor'), choices=lambda: processors.available_processors(), default=processors.get_default_processor)
     urls = TextAreaField(_l('URLs'))
     xlsx_file = FileField(_l('Upload .xlsx file'), validators=[FileAllowed(['xlsx'], _l('Must be .xlsx file!'))])
     file_mapping = SelectField(_l('File mapping'), [validators.DataRequired()], choices={('wachete', 'Wachete mapping'), ('custom','Custom mapping')})
@@ -837,6 +837,8 @@ class processor_text_json_diff_form(commonSettingsForm):
     conditions = FieldList(FormField(ConditionFormRow), min_entries=1)  # Add rule logic here
     use_page_title_in_list = TernaryNoneBooleanField(_l('Use page <title> in list'), default=None)
 
+    history_snapshot_max_length = IntegerField(_l('Number of history items per watch to keep'), render_kw={"style": "width: 5em;"}, validators=[validators.Optional(), validators.NumberRange(min=2)])
+
     def extra_tab_content(self):
         return None
 
@@ -1034,6 +1036,8 @@ class globalSettingsApplicationForm(commonSettingsForm):
                                                                   render_kw={"style": "width: 5em;"},
                                                                   validators=[validators.NumberRange(min=0,
                                                                                                      message=_l("Should contain zero or more attempts"))])
+
+    history_snapshot_max_length = IntegerField(_l('Number of history items per watch to keep'), render_kw={"style": "width: 5em;"}, validators=[validators.Optional(), validators.NumberRange(min=2)])
     ui = FormField(globalSettingsApplicationUIForm)
 
 

changedetectionio/jinja2_custom/safe_jinja.py

@@ -52,7 +52,13 @@ def render(template_str, **args: t.Any) -> str:
     return output[:JINJA2_MAX_RETURN_PAYLOAD_SIZE]
 
 def render_fully_escaped(content):
-    env = jinja2.sandbox.ImmutableSandboxedEnvironment(autoescape=True)
-    template = env.from_string("{{ some_html|e }}")
-    return template.render(some_html=content)
+    """
+    Escape HTML content safely.
+
+    MEMORY LEAK FIX: Use markupsafe.escape() directly instead of creating
+    Jinja2 environments (was causing 1M+ compilations per page load).
+    Simpler, faster, and no concerns about environment state.
+    """
+    from markupsafe import escape
+    return str(escape(content))
 

changedetectionio/model/App.py

@@ -29,7 +29,7 @@ class model(dict):
                     'proxy': None, # Preferred proxy connection
                     'time_between_check': {'weeks': None, 'days': None, 'hours': 3, 'minutes': None, 'seconds': None},
                     'timeout': int(getenv("DEFAULT_SETTINGS_REQUESTS_TIMEOUT", "45")),  # Default 45 seconds
-                    'workers': int(getenv("DEFAULT_SETTINGS_REQUESTS_WORKERS", "10")),  # Number of threads, lower is better for slow connections
+                    'workers': int(getenv("DEFAULT_SETTINGS_REQUESTS_WORKERS", "5")),  # Number of threads, lower is better for slow connections
                     'default_ua': {
                         'html_requests': getenv("DEFAULT_SETTINGS_HEADERS_USERAGENT", DEFAULT_SETTINGS_HEADERS_USERAGENT),
                         'html_webdriver': None,
@@ -46,6 +46,7 @@ class model(dict):
                     'filter_failure_notification_threshold_attempts': _FILTER_FAILURE_THRESHOLD_ATTEMPTS_DEFAULT,
                     'global_ignore_text': [], # List of text to ignore when calculating the comparison checksum
                     'global_subtractive_selectors': [],
+                    'history_snapshot_max_length': None,
                     'ignore_whitespace': True,
                     'ignore_status_codes': False, #@todo implement, as ternary.
                     'ssim_threshold': '0.96',  # Default SSIM threshold for screenshot comparison

changedetectionio/model/Tag.py

@@ -1,10 +1,48 @@
+"""
+Tag/Group domain model for organizing and overriding watch settings.
+
+ARCHITECTURE NOTE: Configuration Override Hierarchy
+===================================================
+
+Tags can override Watch settings when overrides_watch=True.
+Current implementation requires manual checking in processors:
+
+    for tag_uuid in watch.get('tags'):
+        tag = datastore['settings']['application']['tags'][tag_uuid]
+        if tag.get('overrides_watch'):
+            restock_settings = tag.get('restock_settings', {})
+            break
+
+With Pydantic, this would be automatic via chain resolution:
+    Watch → Tag (first with overrides_watch) → Global
+
+See: Watch.py model docstring for full Pydantic architecture explanation
+See: processors/restock_diff/processor.py:184-192 for current manual implementation
+"""
 
 from changedetectionio.model import watch_base
+from changedetectionio.model.persistence import EntityPersistenceMixin
 
+class model(EntityPersistenceMixin, watch_base):
+    """
+    Tag domain model - groups watches and can override their settings.
 
-class model(watch_base):
+    Tags inherit from watch_base to reuse all the same fields as Watch.
+    When overrides_watch=True, tag settings take precedence over watch settings
+    for all watches in this tag/group.
+
+    Fields:
+        overrides_watch (bool): If True, this tag's settings override watch settings
+        title (str): Display name for this tag/group
+        uuid (str): Unique identifier
+        ... (all fields from watch_base can be set as tag-level overrides)
+
+    Resolution order when overrides_watch=True:
+        Watch.field → Tag.field (if overrides_watch) → Global.field
+    """
 
     def __init__(self, *arg, **kw):
+        # Parent class (watch_base) handles __datastore and __datastore_path
         super(model, self).__init__(*arg, **kw)
 
         self['overrides_watch'] = kw.get('default', {}).get('overrides_watch')
@@ -12,3 +50,7 @@ class model(watch_base):
         if kw.get('default'):
             self.update(kw['default'])
             del kw['default']
+
+    # _save_to_disk() method provided by EntityPersistenceMixin
+    # commit() and _get_commit_data() methods inherited from watch_base
+    # Tag uses default _get_commit_data() (includes all keys)

changedetectionio/model/Watch.py

@@ -1,9 +1,37 @@
+"""
+Watch domain model for change detection monitoring.
+
+ARCHITECTURE NOTE: Configuration Override Hierarchy
+===================================================
+
+This module implements Watch objects that inherit from dict (technical debt).
+The dream architecture would use Pydantic for:
+
+1. CHAIN RESOLUTION (Watch → Tag → Global Settings)
+   - Current: Manual resolution scattered across codebase
+   - Future: @computed_field properties with automatic resolution
+   - Examples: resolved_fetch_backend, resolved_restock_settings, etc.
+
+2. DATABASE BACKEND ABSTRACTION
+   - Current: Domain model tightly coupled to file-based JSON storage
+   - Future: Domain model (Pydantic) separate from persistence layer
+   - Enables: Easy migration to PostgreSQL, MongoDB, etc.
+
+3. TYPE SAFETY & VALIDATION
+   - Current: Dict access with no compile-time checks
+   - Future: Type hints, IDE autocomplete, validation at boundaries
+
+See class model docstring for detailed explanation and examples.
+See: processors/restock_diff/processor.py:184-192 for manual resolution example
+"""
+
 from blinker import signal
 from changedetectionio.validate_url import is_safe_valid_url
 
 from changedetectionio.strtobool import strtobool
 from changedetectionio.jinja2_custom import render as jinja_render
 from . import watch_base
+from .persistence import EntityPersistenceMixin
 import os
 import re
 from pathlib import Path
@@ -13,15 +41,16 @@ from .. import jinja2_custom as safe_jinja
 from ..html_tools import TRANSLATE_WHITESPACE_TABLE
 
 FAVICON_RESAVE_THRESHOLD_SECONDS=86400
-BROTLI_COMPRESS_SIZE_THRESHOLD = int(os.getenv('SNAPSHOT_BROTLI_COMPRESSION_THRESHOLD', 1024))
+BROTLI_COMPRESS_SIZE_THRESHOLD = int(os.getenv('SNAPSHOT_BROTLI_COMPRESSION_THRESHOLD', 1024*20))
 
 minimum_seconds_recheck_time = int(os.getenv('MINIMUM_SECONDS_RECHECK_TIME', 3))
 mtable = {'seconds': 1, 'minutes': 60, 'hours': 3600, 'days': 86400, 'weeks': 86400 * 7}
 
 def _brotli_save(contents, filepath, mode=None, fallback_uncompressed=False):
     """
-    Save compressed data using native brotli.
-    Testing shows no memory leak when using gc.collect() after compression.
+    Save compressed data using native brotli with streaming compression.
+    Uses chunked compression to minimize peak memory usage and malloc_trim()
+    to force release of C-level memory back to the OS.
 
     Args:
         contents: data to compress (str or bytes)
@@ -37,27 +66,52 @@ def _brotli_save(contents, filepath, mode=None, fallback_uncompressed=False):
     """
     import brotli
     import gc
+    import ctypes
 
     # Ensure contents are bytes
     if isinstance(contents, str):
         contents = contents.encode('utf-8')
 
     try:
-        logger.debug(f"Starting brotli compression of {len(contents)} bytes.")
+        original_size = len(contents)
+        logger.debug(f"Starting brotli streaming compression of {original_size} bytes.")
 
-        if mode is not None:
-            compressed_data = brotli.compress(contents, mode=mode)
-        else:
-            compressed_data = brotli.compress(contents)
+        # Create streaming compressor
+        compressor = brotli.Compressor(quality=6, mode=mode if mode is not None else brotli.MODE_GENERIC)
+
+        # Stream compress in chunks to minimize memory usage
+        chunk_size = 65536  # 64KB chunks
+        total_compressed_size = 0
 
         with open(filepath, 'wb') as f:
-            f.write(compressed_data)
+            # Process data in chunks
+            offset = 0
+            while offset < len(contents):
+                chunk = contents[offset:offset + chunk_size]
+                compressed_chunk = compressor.process(chunk)
+                if compressed_chunk:
+                    f.write(compressed_chunk)
+                    total_compressed_size += len(compressed_chunk)
+                offset += chunk_size
 
-        logger.debug(f"Finished brotli compression - From {len(contents)} to {len(compressed_data)} bytes.")
+            # Finalize compression - critical for proper cleanup
+            final_chunk = compressor.finish()
+            if final_chunk:
+                f.write(final_chunk)
+                total_compressed_size += len(final_chunk)
 
-        # Force garbage collection to prevent memory buildup
+        logger.debug(f"Finished brotli compression - From {original_size} to {total_compressed_size} bytes.")
+
+        # Cleanup: Delete compressor, force Python GC, then force C-level memory release
+        del compressor
         gc.collect()
 
+        # Force release of C-level memory back to OS (since brotli is a C library)
+        try:
+            ctypes.CDLL('libc.so.6').malloc_trim(0)
+        except Exception:
+            pass  # malloc_trim not available on all systems (e.g., macOS)
+
         return filepath
 
     except Exception as e:
@@ -74,17 +128,112 @@ def _brotli_save(contents, filepath, mode=None, fallback_uncompressed=False):
             raise Exception(f"Brotli compression failed for {filepath}: {e}")
 
 
-class model(watch_base):
+class model(EntityPersistenceMixin, watch_base):
+    """
+    Watch domain model for monitoring URL changes.
+
+    Inherits from watch_base (which inherits dict) - see watch_base docstring for field documentation.
+
+    ## Configuration Override Hierarchy (Chain Resolution)
+
+    The dream architecture uses a 3-level resolution chain:
+        Watch settings → Tag/Group settings → Global settings
+
+    Current implementation is MANUAL (see processor.py:184-192 for example):
+        - Processors manually check watch.get('field')
+        - Then loop through watch.tags to find first tag with overrides_watch=True
+        - Finally fall back to datastore['settings']['application']['field']
+
+    FUTURE: Pydantic-based chain resolution would enable:
+
+        ```python
+        # Instead of manual resolution in every processor:
+        restock_settings = watch.get('restock_settings', {})
+        for tag_uuid in watch.get('tags'):
+            tag = datastore['settings']['application']['tags'][tag_uuid]
+            if tag.get('overrides_watch'):
+                restock_settings = tag.get('restock_settings', {})
+                break
+
+        # Clean computed properties with automatic resolution:
+        @computed_field
+        def resolved_restock_settings(self) -> dict:
+            if self.restock_settings:
+                return self.restock_settings
+            for tag_uuid in self.tags:
+                tag = self._datastore.get_tag(tag_uuid)
+                if tag.overrides_watch and tag.restock_settings:
+                    return tag.restock_settings
+            return self._datastore.settings.restock_settings or {}
+
+        # Usage: watch.resolved_restock_settings (automatic, type-safe, tested once)
+        ```
+
+    Benefits of Pydantic migration:
+        1. Single source of truth for resolution logic (not scattered across processors)
+        2. Type safety + IDE autocomplete (watch.resolved_fetch_backend vs dict navigation)
+        3. Database backend abstraction (domain model separate from persistence)
+        4. Automatic validation at boundaries
+        5. Self-documenting via type hints
+        6. Easy to test resolution independently
+
+    Resolution chain examples that would benefit:
+        - fetch_backend: watch → tag → global (see get_fetch_backend property)
+        - notification_urls: watch → tag → global
+        - time_between_check: watch → global (see threshold_seconds)
+        - restock_settings: watch → tag (see processors/restock_diff/processor.py:184-192)
+        - history_snapshot_max_length: watch → global (see save_history_blob:550-556)
+        - All processor_config_* settings could use tag overrides
+
+    ## Database Backend Abstraction with Pydantic
+
+    Current: Watch inherits dict, tightly coupled to file-based JSON storage
+    Future: Domain model (Watch) separate from persistence layer
+
+        ```python
+        # Domain model (database-agnostic)
+        class Watch(BaseModel):
+            uuid: str
+            url: str
+            # ... validation, business logic
+
+        # Pluggable backends
+        class DataStoreBackend(ABC):
+            def save_watch(self, watch: Watch): ...
+            def load_watch(self, uuid: str) -> Watch: ...
+
+        # Implementations: FileBackend, MongoBackend, PostgresBackend, etc.
+        ```
+
+    This would enable:
+        - Easy migration between storage backends (file → postgres → mongodb)
+        - Pydantic handles serialization/deserialization automatically
+        - Domain logic stays clean (no storage concerns in Watch methods)
+
+    ## Migration Path
+
+    Given existing codebase, incremental migration recommended:
+        1. Create Pydantic models alongside existing dict-based models
+        2. Add .to_pydantic() / .from_pydantic() bridge methods
+        3. Gradually migrate code to use Pydantic models
+        4. Remove dict inheritance once migration complete
+
+    See: watch_base docstring for technical debt discussion
+    See: processors/restock_diff/processor.py:184-192 for manual resolution example
+    See: Watch.py:550-556 for nested dict navigation that would become watch.resolved_*
+    """
     __newest_history_key = None
     __history_n = 0
     jitter_seconds = 0
 
     def __init__(self, *arg, **kw):
-        self.__datastore_path = kw.get('datastore_path')
-        if kw.get('datastore_path'):
-            del kw['datastore_path']
-            
+        # Validate __datastore before calling parent (Watch requires it)
+        if not kw.get('__datastore'):
+            raise ValueError("Watch object requires '__datastore' reference - cannot access global settings without it")
+
+        # Parent class (watch_base) handles __datastore and __datastore_path
         super(model, self).__init__(*arg, **kw)
+
         if kw.get('default'):
             self.update(kw['default'])
             del kw['default']
@@ -95,6 +244,9 @@ class model(watch_base):
         # Be sure the cached timestamp is ready
         bump = self.history
 
+    # Note: __deepcopy__, __getstate__, and __setstate__ are inherited from watch_base
+    # This prevents memory leaks by sharing __datastore reference instead of copying it
+
     @property
     def viewed(self):
         # Don't return viewed when last_viewed is 0 and newest_key is 0
@@ -107,11 +259,6 @@ class model(watch_base):
     def has_unviewed(self):
         return int(self.newest_history_key) > int(self['last_viewed']) and self.__history_n >= 2
 
-    def ensure_data_dir_exists(self):
-        if not os.path.isdir(self.watch_data_dir):
-            logger.debug(f"> Creating data dir {self.watch_data_dir}")
-            os.mkdir(self.watch_data_dir)
-
     @property
     def link(self):
 
@@ -167,7 +314,8 @@ class model(watch_base):
 
         # JSON Data, Screenshots, Textfiles (history index and snapshots), HTML in the future etc
         # But preserve processor config files (they're configuration, not history data)
-        for item in pathlib.Path(str(self.watch_data_dir)).rglob("*.*"):
+        # Use glob not rglob here for safety.
+        for item in pathlib.Path(str(self.data_dir)).glob("*.*"):
             # Skip processor config files
             if item.name in processor_config_files:
                 continue
@@ -204,8 +352,30 @@ class model(watch_base):
     @property
     def get_fetch_backend(self):
         """
-        Like just using the `fetch_backend` key but there could be some logic
-        :return:
+        Get the fetch backend for this watch with special case handling.
+
+        CHAIN RESOLUTION OPPORTUNITY:
+        Currently returns watch.fetch_backend directly, but doesn't implement
+        Watch → Tag → Global resolution chain. With Pydantic:
+
+        @computed_field
+        def resolved_fetch_backend(self) -> str:
+            # Special case: PDFs always use html_requests
+            if self.is_pdf:
+                return 'html_requests'
+
+            # Watch override
+            if self.fetch_backend and self.fetch_backend != 'system':
+                return self.fetch_backend
+
+            # Tag override (first tag with overrides_watch=True wins)
+            for tag_uuid in self.tags:
+                tag = self._datastore.get_tag(tag_uuid)
+                if tag.overrides_watch and tag.fetch_backend:
+                    return tag.fetch_backend
+
+            # Global default
+            return self._datastore.settings.fetch_backend
         """
         # Maybe also if is_image etc?
         # This is because chrome/playwright wont render the PDF in the browser and we will just fetch it and use pdf2html to see the text.
@@ -254,11 +424,11 @@ class model(watch_base):
         tmp_history = {}
 
         # In the case we are only using the watch for processing without history
-        if not self.watch_data_dir:
+        if not self.data_dir:
             return []
 
         # Read the history file as a dict
-        fname = os.path.join(self.watch_data_dir, self.history_index_filename)
+        fname = os.path.join(self.data_dir, self.history_index_filename)
         if os.path.isfile(fname):
             logger.debug(f"Reading watch history index for {self.get('uuid')}")
             with open(fname, "r", encoding='utf-8') as f:
@@ -271,13 +441,13 @@ class model(watch_base):
                         # Cross-platform: check for any path separator (works on Windows and Unix)
                         if os.sep not in v and '/' not in v and '\\' not in v:
                             # Relative filename only, no path separators
-                            v = os.path.join(self.watch_data_dir, v)
+                            v = os.path.join(self.data_dir, v)
                         else:
                             # It's possible that they moved the datadir on older versions
                             # So the snapshot exists but is in a different path
                             # Cross-platform: use os.path.basename instead of split('/')
                             snapshot_fname = os.path.basename(v)
-                            proposed_new_path = os.path.join(self.watch_data_dir, snapshot_fname)
+                            proposed_new_path = os.path.join(self.data_dir, snapshot_fname)
                             if not os.path.exists(v) and os.path.exists(proposed_new_path):
                                 v = proposed_new_path
 
@@ -294,7 +464,7 @@ class model(watch_base):
 
     @property
     def has_history(self):
-        fname = os.path.join(self.watch_data_dir, self.history_index_filename)
+        fname = os.path.join(self.data_dir, self.history_index_filename)
         return os.path.isfile(fname)
 
     @property
@@ -397,16 +567,49 @@ class model(watch_base):
         with open(filepath, 'r', encoding='utf-8', errors='ignore') as f:
             return f.read()
 
-    def _write_atomic(self, dest, data):
+    def _write_atomic(self, dest, data, mode='wb'):
         """Write data atomically to dest using a temp file"""
-        if not os.path.exists(dest):
-            import tempfile
-            with tempfile.NamedTemporaryFile('wb', delete=False, dir=self.watch_data_dir) as tmp:
-                tmp.write(data)
-                tmp.flush()
-                os.fsync(tmp.fileno())
-                tmp_path = tmp.name
-            os.replace(tmp_path, dest)
+        import tempfile
+        with tempfile.NamedTemporaryFile(mode, delete=False, dir=self.data_dir) as tmp:
+            tmp.write(data)
+            tmp.flush()
+            os.fsync(tmp.fileno())
+            tmp_path = tmp.name
+        os.replace(tmp_path, dest)
+
+    def history_trim(self, newest_n_items):
+        from pathlib import Path
+        import gc
+        # Sort by timestamp (key)
+        sorted_items = sorted(self.history.items(), key=lambda x: int(x[0]))
+
+        keep_part = dict(sorted_items[-newest_n_items:])
+        delete_part = dict(sorted_items[:-newest_n_items])
+        logger.info( f"[{self.get('uuid')}] Trimming history to most recent {newest_n_items} items, keeping {len(keep_part)} items deleting {len(delete_part)} items.")
+
+        if delete_part:
+            for item in delete_part.items():
+                try:
+                    Path(item[1]).unlink(missing_ok=True)
+                except Exception as e:
+                    logger.critical(f"{str(e)}")
+                finally:
+                    logger.debug(f"[{self.get('uuid')}] Deleted {item[1]} history snapshot")
+        try:
+            dest = os.path.join(self.data_dir, self.history_index_filename)
+            output = "\r\n".join(
+                f"{k},{Path(v).name}"
+                for k, v in keep_part.items()
+            )+"\r\n"
+            self._write_atomic(dest=dest, data=output, mode='w')
+        except Exception as e:
+            logger.critical(f"{str(e)}")
+        finally:
+            logger.debug(f"[{self.get('uuid')}] Updated history index {dest}")
+
+        # reimport
+        bump = self.history
+        gc.collect()
 
     # Save some text file to the appropriate path and bump the history
     # result_obj from fetch_site_status.run()
@@ -415,7 +618,6 @@ class model(watch_base):
         logger.trace(f"{self.get('uuid')} - Updating {self.history_index_filename} with timestamp {timestamp}")
 
         self.ensure_data_dir_exists()
-
         skip_brotli = strtobool(os.getenv('DISABLE_BROTLI_TEXT_SNAPSHOT', 'False'))
 
         # Binary data - detect file type and save without compression
@@ -433,7 +635,7 @@ class model(watch_base):
                 ext = 'bin'
 
             snapshot_fname = f"{snapshot_id}.{ext}"
-            dest = os.path.join(self.watch_data_dir, snapshot_fname)
+            dest = os.path.join(self.data_dir, snapshot_fname)
             self._write_atomic(dest, contents)
             logger.trace(f"Saved binary snapshot as {snapshot_fname} ({len(contents)} bytes)")
 
@@ -443,7 +645,7 @@ class model(watch_base):
                 # Compressed text
                 import brotli
                 snapshot_fname = f"{snapshot_id}.txt.br"
-                dest = os.path.join(self.watch_data_dir, snapshot_fname)
+                dest = os.path.join(self.data_dir, snapshot_fname)
 
                 if not os.path.exists(dest):
                     try:
@@ -454,16 +656,16 @@ class model(watch_base):
                         logger.error(f"{self.get('uuid')} - Brotli compression failed: {e}")
                         # Fallback to uncompressed
                         snapshot_fname = f"{snapshot_id}.txt"
-                        dest = os.path.join(self.watch_data_dir, snapshot_fname)
+                        dest = os.path.join(self.data_dir, snapshot_fname)
                         self._write_atomic(dest, contents.encode('utf-8'))
             else:
                 # Plain text
                 snapshot_fname = f"{snapshot_id}.txt"
-                dest = os.path.join(self.watch_data_dir, snapshot_fname)
+                dest = os.path.join(self.data_dir, snapshot_fname)
                 self._write_atomic(dest, contents.encode('utf-8'))
 
         # Append to history.txt atomically
-        index_fname = os.path.join(self.watch_data_dir, self.history_index_filename)
+        index_fname = os.path.join(self.data_dir, self.history_index_filename)
         index_line = f"{timestamp},{snapshot_fname}\n"
 
         with open(index_fname, 'a', encoding='utf-8') as f:
@@ -475,6 +677,17 @@ class model(watch_base):
         self.__newest_history_key = timestamp
         self.__history_n += 1
 
+        # MANUAL CHAIN RESOLUTION: Watch → Global
+        # With Pydantic, this would become: maxlen = watch.resolved_history_snapshot_max_length
+        # @computed_field def resolved_history_snapshot_max_length(self) -> Optional[int]:
+        #     if self.history_snapshot_max_length: return self.history_snapshot_max_length
+        #     if tag := self._get_override_tag(): return tag.history_snapshot_max_length
+        #     return self._datastore.settings.history_snapshot_max_length
+        maxlen = self.get('history_snapshot_max_length') or self.get_global_setting('application', 'history_snapshot_max_length')
+
+        if maxlen and self.__history_n and self.__history_n > maxlen:
+            self.history_trim(newest_n_items=maxlen)
+
         # @todo bump static cache of the last timestamp so we dont need to examine the file to set a proper ''viewed'' status
         return snapshot_fname
 
@@ -527,7 +740,7 @@ class model(watch_base):
         return not local_lines.issubset(existing_history)
 
     def get_screenshot(self):
-        fname = os.path.join(self.watch_data_dir, "last-screenshot.png")
+        fname = os.path.join(self.data_dir, "last-screenshot.png")
         if os.path.isfile(fname):
             return fname
 
@@ -542,7 +755,7 @@ class model(watch_base):
         if not favicon_fname:
             return True
         try:
-            fname = next(iter(glob.glob(os.path.join(self.watch_data_dir, "favicon.*"))), None)
+            fname = next(iter(glob.glob(os.path.join(self.data_dir, "favicon.*"))), None)
             logger.trace(f"Favicon file maybe found at {fname}")
             if os.path.isfile(fname):
                 file_age = int(time.time() - os.path.getmtime(fname))
@@ -575,7 +788,7 @@ class model(watch_base):
             base = "favicon"
             extension = "ico"
 
-        fname = os.path.join(self.watch_data_dir, f"favicon.{extension}")
+        fname = os.path.join(self.data_dir, f"favicon.{extension}")
 
         try:
             # validate=True makes sure the string only contains valid base64 chars
@@ -587,6 +800,11 @@ class model(watch_base):
                 try:
                     with open(fname, 'wb') as f:
                         f.write(decoded)
+
+                    # Invalidate favicon filename cache
+                    if hasattr(self, '_favicon_filename_cache'):
+                        delattr(self, '_favicon_filename_cache')
+
                     # A signal that could trigger the socket server to update the browser also
                     watch_check_update = signal('watch_favicon_bump')
                     if watch_check_update:
@@ -603,20 +821,32 @@ class model(watch_base):
         Find any favicon.* file in the current working directory
         and return the contents of the newest one.
 
+        MEMORY LEAK FIX: Cache the result to avoid repeated glob.glob() operations.
+        glob.glob() causes millions of fnmatch allocations when called for every watch on page load.
+
         Returns:
-            bytes: Contents of the newest favicon file, or None if not found.
+            str: Basename of the newest favicon file, or None if not found.
         """
+        # Check cache first (prevents 26M+ allocations from repeated glob operations)
+        cache_key = '_favicon_filename_cache'
+        if hasattr(self, cache_key):
+            return getattr(self, cache_key)
+
         import glob
 
         # Search for all favicon.* files
-        files = glob.glob(os.path.join(self.watch_data_dir, "favicon.*"))
+        files = glob.glob(os.path.join(self.data_dir, "favicon.*"))
 
         if not files:
-            return None
+            result = None
+        else:
+            # Find the newest by modification time
+            newest_file = max(files, key=os.path.getmtime)
+            result = os.path.basename(newest_file)
 
-        # Find the newest by modification time
-        newest_file = max(files, key=os.path.getmtime)
-        return os.path.basename(newest_file)
+        # Cache the result
+        setattr(self, cache_key, result)
+        return result
 
     def get_screenshot_as_thumbnail(self, max_age=3200):
         """Return path to a square thumbnail of the most recent screenshot.
@@ -632,7 +862,7 @@ class model(watch_base):
         import os
         import time
 
-        thumbnail_path = os.path.join(self.watch_data_dir, "thumbnail.jpeg")
+        thumbnail_path = os.path.join(self.data_dir, "thumbnail.jpeg")
         top_trim = 500  # Pixels from top of screenshot to use
 
         screenshot_path = self.get_screenshot()
@@ -683,7 +913,7 @@ class model(watch_base):
             return None
 
     def __get_file_ctime(self, filename):
-        fname = os.path.join(self.watch_data_dir, filename)
+        fname = os.path.join(self.data_dir, filename)
         if os.path.isfile(fname):
             return int(os.path.getmtime(fname))
         return False
@@ -708,14 +938,9 @@ class model(watch_base):
     def snapshot_error_screenshot_ctime(self):
         return self.__get_file_ctime('last-error-screenshot.png')
 
-    @property
-    def watch_data_dir(self):
-        # The base dir of the watch data
-        return os.path.join(self.__datastore_path, self['uuid']) if self.__datastore_path else None
-
     def get_error_text(self):
         """Return the text saved from a previous request that resulted in a non-200 error"""
-        fname = os.path.join(self.watch_data_dir, "last-error.txt")
+        fname = os.path.join(self.data_dir, "last-error.txt")
         if os.path.isfile(fname):
             with open(fname, 'r', encoding='utf-8') as f:
                 return f.read()
@@ -723,7 +948,7 @@ class model(watch_base):
 
     def get_error_snapshot(self):
         """Return path to the screenshot that resulted in a non-200 error"""
-        fname = os.path.join(self.watch_data_dir, "last-error-screenshot.png")
+        fname = os.path.join(self.data_dir, "last-error-screenshot.png")
         if os.path.isfile(fname):
             return fname
         return False
@@ -747,6 +972,37 @@ class model(watch_base):
     def toggle_mute(self):
         self['notification_muted'] ^= True
 
+    def _get_commit_data(self):
+        """
+        Prepare watch data for commit.
+
+        Excludes processor_config_* keys (stored in separate files).
+        Normalizes browser_steps to empty list if no meaningful steps.
+        """
+        import copy
+
+        # Get base snapshot with lock
+        lock = self._datastore.lock if self._datastore and hasattr(self._datastore, 'lock') else None
+
+        if lock:
+            with lock:
+                snapshot = dict(self)
+        else:
+            snapshot = dict(self)
+
+        # Exclude processor config keys (stored separately)
+        watch_dict = {k: copy.deepcopy(v) for k, v in snapshot.items() if not k.startswith('processor_config_')}
+
+        # Normalize browser_steps: if no meaningful steps, save as empty list
+        if not self.has_browser_steps:
+            watch_dict['browser_steps'] = []
+
+        return watch_dict
+
+    # _save_to_disk() method provided by EntityPersistenceMixin
+    # commit() method inherited from watch_base
+
+
     def extra_notification_token_values(self):
         # Used for providing extra tokens
         # return {'widget': 555}
@@ -776,7 +1032,7 @@ class model(watch_base):
                         if not csv_writer:
                             # A file on the disk can be transferred much faster via flask than a string reply
                             csv_output_filename = f"report-{self.get('uuid')}.csv"
-                            f = open(os.path.join(self.watch_data_dir, csv_output_filename), 'w')
+                            f = open(os.path.join(self.data_dir, csv_output_filename), 'w')
                             # @todo some headers in the future
                             #fieldnames = ['Epoch seconds', 'Date']
                             csv_writer = csv.writer(f,
@@ -818,7 +1074,7 @@ class model(watch_base):
 
     def save_error_text(self, contents):
         self.ensure_data_dir_exists()
-        target_path = os.path.join(self.watch_data_dir, "last-error.txt")
+        target_path = os.path.join(self.data_dir, "last-error.txt")
         with open(target_path, 'w', encoding='utf-8') as f:
             f.write(contents)
 
@@ -827,9 +1083,9 @@ class model(watch_base):
         import zlib
 
         if as_error:
-            target_path = os.path.join(str(self.watch_data_dir), "elements-error.deflate")
+            target_path = os.path.join(str(self.data_dir), "elements-error.deflate")
         else:
-            target_path = os.path.join(str(self.watch_data_dir), "elements.deflate")
+            target_path = os.path.join(str(self.data_dir), "elements.deflate")
 
         self.ensure_data_dir_exists()
 
@@ -844,9 +1100,9 @@ class model(watch_base):
     def save_screenshot(self, screenshot: bytes, as_error=False):
 
         if as_error:
-            target_path = os.path.join(self.watch_data_dir, "last-error-screenshot.png")
+            target_path = os.path.join(self.data_dir, "last-error-screenshot.png")
         else:
-            target_path = os.path.join(self.watch_data_dir, "last-screenshot.png")
+            target_path = os.path.join(self.data_dir, "last-screenshot.png")
 
         self.ensure_data_dir_exists()
 
@@ -857,7 +1113,7 @@ class model(watch_base):
 
     def get_last_fetched_text_before_filters(self):
         import brotli
-        filepath = os.path.join(self.watch_data_dir, 'last-fetched.br')
+        filepath = os.path.join(self.data_dir, 'last-fetched.br')
 
         if not os.path.isfile(filepath) or os.path.getsize(filepath) == 0:
             # If a previous attempt doesnt yet exist, just snarf the previous snapshot instead
@@ -872,13 +1128,13 @@ class model(watch_base):
 
     def save_last_text_fetched_before_filters(self, contents):
         import brotli
-        filepath = os.path.join(self.watch_data_dir, 'last-fetched.br')
+        filepath = os.path.join(self.data_dir, 'last-fetched.br')
         _brotli_save(contents, filepath, mode=brotli.MODE_TEXT, fallback_uncompressed=False)
 
     def save_last_fetched_html(self, timestamp, contents):
         self.ensure_data_dir_exists()
         snapshot_fname = f"{timestamp}.html.br"
-        filepath = os.path.join(self.watch_data_dir, snapshot_fname)
+        filepath = os.path.join(self.data_dir, snapshot_fname)
         _brotli_save(contents, filepath, mode=None, fallback_uncompressed=True)
         self._prune_last_fetched_html_snapshots()
 
@@ -886,7 +1142,7 @@ class model(watch_base):
         import brotli
 
         snapshot_fname = f"{timestamp}.html.br"
-        filepath = os.path.join(self.watch_data_dir, snapshot_fname)
+        filepath = os.path.join(self.data_dir, snapshot_fname)
         if os.path.isfile(filepath):
             with open(filepath, 'rb') as f:
                 return (brotli.decompress(f.read()).decode('utf-8'))
@@ -901,7 +1157,7 @@ class model(watch_base):
 
         for index, timestamp in enumerate(dates):
             snapshot_fname = f"{timestamp}.html.br"
-            filepath = os.path.join(self.watch_data_dir, snapshot_fname)
+            filepath = os.path.join(self.data_dir, snapshot_fname)
 
             # Keep only the first 2
             if index > 1 and os.path.isfile(filepath):
@@ -912,7 +1168,7 @@ class model(watch_base):
     def get_browsersteps_available_screenshots(self):
         "For knowing which screenshots are available to show the user in BrowserSteps UI"
         available = []
-        for f in Path(self.watch_data_dir).glob('step_before-*.jpeg'):
+        for f in Path(self.data_dir).glob('step_before-*.jpeg'):
             step_n=re.search(r'step_before-(\d+)', f.name)
             if step_n:
                 available.append(step_n.group(1))

changedetectionio/model/__init__.py

@@ -2,12 +2,169 @@ import os
 import uuid
 
 from changedetectionio import strtobool
+from .persistence import EntityPersistenceMixin, _determine_entity_type
+
+__all__ = ['EntityPersistenceMixin', 'watch_base']
+
 USE_SYSTEM_DEFAULT_NOTIFICATION_FORMAT_FOR_WATCH = 'System default'
 CONDITIONS_MATCH_LOGIC_DEFAULT = 'ALL'
 
+
 class watch_base(dict):
+    """
+    Base watch domain model (inherits from dict for backward compatibility).
+
+    WARNING: This class inherits from dict, which violates proper encapsulation.
+    Dict inheritance is legacy technical debt that should be refactored to a proper
+    domain model (e.g., Pydantic BaseModel) for better type safety and validation.
+
+    TODO: Migrate to Pydantic BaseModel for:
+          - Type safety and IDE autocomplete
+          - Automatic validation
+          - Clear separation between domain model and serialization
+          - Database backend abstraction (file → postgres → mongodb)
+          - Configuration override chain resolution (Watch → Tag → Global)
+          - Immutability options
+          - Better testing
+
+    CHAIN RESOLUTION ARCHITECTURE:
+        The dream is a 3-level override hierarchy:
+            Watch settings → Tag/Group settings → Global settings
+
+        Current implementation: MANUAL resolution scattered across codebase
+        - Processors manually check watch.get('field')
+        - Loop through tags to find overrides_watch=True
+        - Fall back to datastore['settings']['application']['field']
+
+        Pydantic implementation: AUTOMATIC resolution via @computed_field
+        - Single source of truth for each setting's resolution logic
+        - Type-safe, testable, self-documenting
+        - Example: watch.resolved_fetch_backend (instead of nested dict navigation)
+
+        See: Watch.py model docstring for detailed Pydantic architecture plan
+        See: Tag.py model docstring for tag override explanation
+        See: processors/restock_diff/processor.py:184-192 for current manual example
+
+    Core Fields:
+        uuid (str): Unique identifier for this watch (auto-generated)
+        url (str): Target URL to monitor for changes
+        title (str|None): Custom display name (overrides page_title if set)
+        page_title (str|None): Title extracted from <title> tag of monitored page
+        tags (List[str]): List of tag UUIDs for categorization
+        tag (str): DEPRECATED - Old single-tag system, use tags instead
+
+    Check Configuration:
+        processor (str): Processor type ('text_json_diff', 'restock_diff', etc.)
+        fetch_backend (str): Fetcher to use ('system', 'html_requests', 'playwright', etc.)
+        method (str): HTTP method ('GET', 'POST', etc.)
+        headers (dict): Custom HTTP headers to send
+        proxy (str|None): Preferred proxy server
+        paused (bool): Whether change detection is paused
+
+    Scheduling:
+        time_between_check (dict): Check interval {'weeks': int, 'days': int, 'hours': int, 'minutes': int, 'seconds': int}
+        time_between_check_use_default (bool): Use global default interval if True
+        time_schedule_limit (dict): Weekly schedule limiting when checks can run
+            Structure: {
+                'enabled': bool,
+                'monday/tuesday/.../sunday': {
+                    'enabled': bool,
+                    'start_time': str ('HH:MM'),
+                    'duration': {'hours': str, 'minutes': str}
+                }
+            }
+
+    Content Filtering:
+        include_filters (List[str]): CSS/XPath selectors to extract content
+        subtractive_selectors (List[str]): Selectors to remove from content
+        ignore_text (List[str]): Text patterns to ignore in change detection
+        trigger_text (List[str]): Text/regex that must be present to trigger change
+        text_should_not_be_present (List[str]): Text that should NOT be present
+        extract_text (List[str]): Regex patterns to extract specific text after filtering
+
+    Text Processing:
+        trim_text_whitespace (bool): Strip leading/trailing whitespace
+        sort_text_alphabetically (bool): Sort lines alphabetically before comparison
+        remove_duplicate_lines (bool): Remove duplicate lines
+        check_unique_lines (bool): Compare against all history for unique lines
+        strip_ignored_lines (bool|None): Remove lines matching ignore patterns
+
+    Change Detection Filters:
+        filter_text_added (bool): Include added text in change detection
+        filter_text_removed (bool): Include removed text in change detection
+        filter_text_replaced (bool): Include replaced text in change detection
+
+    Browser Automation:
+        browser_steps (List[dict]): Browser automation steps for JS-heavy sites
+        browser_steps_last_error_step (int|None): Last step that caused error
+        webdriver_delay (int|None): Seconds to wait after page load
+        webdriver_js_execute_code (str|None): JavaScript to execute before extraction
+
+    Restock Detection:
+        in_stock_only (bool): Only trigger on in-stock transitions
+        follow_price_changes (bool): Monitor price changes
+        has_ldjson_price_data (bool|None): Whether page has LD-JSON price data
+        track_ldjson_price_data (str|None): Track LD-JSON price data ('ACCEPT', 'REJECT', None)
+        price_change_threshold_percent (float|None): Minimum price change % to trigger
+
+    Notifications:
+        notification_urls (List[str]): Apprise URLs for notifications
+        notification_title (str|None): Custom notification title template
+        notification_body (str|None): Custom notification body template
+        notification_format (str): Notification format (e.g., 'System default', 'Text', 'HTML')
+        notification_muted (bool): Disable notifications for this watch
+        notification_screenshot (bool): Include screenshot in notifications
+        notification_alert_count (int): Number of notifications sent
+        last_notification_error (str|None): Last notification error message
+        body (str|None): DEPRECATED? Legacy notification body field
+        filter_failure_notification_send (bool): Send notification on filter failures
+
+    History & State:
+        date_created (int|None): Unix timestamp of watch creation
+        last_checked (int): Unix timestamp of last check
+        last_viewed (int): History snapshot key of last user view
+        last_error (str|bool): Last error message or False if no error
+        check_count (int): Total number of checks performed
+        fetch_time (float): Duration of last fetch in seconds
+        consecutive_filter_failures (int): Counter for consecutive filter match failures
+        previous_md5 (str|bool): MD5 hash of previous content
+        previous_md5_before_filters (str|bool): MD5 hash before filters applied
+        history_snapshot_max_length (int|None): Max history snapshots to keep (None = use global)
+
+    Conditions:
+        conditions (dict): Custom conditions for change detection logic
+        conditions_match_logic (str): Logic operator ('ALL', 'ANY') for conditions
+
+    Metadata:
+        content-type (str|None): Content-Type from last fetch
+        remote_server_reply (str|None): Server header from last response
+        ignore_status_codes (List[int]|None): HTTP status codes to ignore
+        use_page_title_in_list (bool|None): Display page title in watch list (None = use system default)
+
+    Instance Attributes (not serialized):
+        __datastore: Reference to parent DataStore (set externally after creation)
+        data_dir: Filesystem path for this watch's data directory
+
+    Notes:
+        - Many fields default to None to distinguish "not set" from "set to default"
+        - When field is None, system-level defaults are used
+        - Processor-specific configs (e.g., processor_config_*) are NOT stored in watch.json
+          They are stored in separate {processor_name}.json files
+        - This class is used for both Watch and Tag objects (tags reuse the structure)
+    """
 
     def __init__(self, *arg, **kw):
+        # Store datastore reference (common to Watch and Tag)
+        # Use single underscore to avoid name mangling issues in subclasses
+        self._datastore = kw.get('__datastore')
+        if kw.get('__datastore'):
+            del kw['__datastore']
+
+        # Store datastore_path (common to Watch and Tag)
+        self._datastore_path = kw.get('datastore_path')
+        if kw.get('datastore_path'):
+            del kw['datastore_path']
+
         self.update({
             # Custom notification content
             # Re #110, so then if this is set to None, we know to use the default value instead
@@ -32,6 +189,7 @@ class watch_base(dict):
             'filter_text_replaced': True,
             'follow_price_changes': True,
             'has_ldjson_price_data': None,
+            'history_snapshot_max_length': None,
             'headers': {},  # Extra headers to send
             'ignore_text': [],  # List of text to ignore when calculating the comparison checksum
             'ignore_status_codes': None,
@@ -139,4 +297,223 @@ class watch_base(dict):
         super(watch_base, self).__init__(*arg, **kw)
 
         if self.get('default'):
-            del self['default']
+            del self['default']
+
+    def __deepcopy__(self, memo):
+        """
+        Custom deepcopy for all watch_base subclasses (Watch, Tag, etc.).
+
+        CRITICAL FIX: Prevents copying large reference objects like __datastore
+        which would cause exponential memory growth when Watch objects are deepcopied.
+
+        This is called by:
+        - api/Watch.py:76 (API endpoint)
+        - api/Tags.py:28 (Tags API)
+        - processors/base.py:26 (EVERY processor run)
+        - store/__init__.py:544 (clone watch)
+        - And other locations
+        """
+        from copy import deepcopy
+
+        # Create new instance without calling __init__
+        cls = self.__class__
+        new_obj = cls.__new__(cls)
+        memo[id(self)] = new_obj
+
+        # Copy the dict data (all the settings)
+        for key, value in self.items():
+            new_obj[key] = deepcopy(value, memo)
+
+        # Copy instance attributes dynamically
+        # This handles Watch-specific attrs (like __datastore) and any future subclass attrs
+        for attr_name in dir(self):
+            # Skip methods, special attrs, and dict keys
+            if attr_name.startswith('_') and not attr_name.startswith('__'):
+                # This catches _model__datastore, _model__history_n, etc.
+                try:
+                    attr_value = getattr(self, attr_name)
+
+                    # Special handling: Share references to large objects instead of copying
+                    # Examples: _datastore, __datastore, __app_reference, __global_settings, etc.
+                    if (attr_name == '_datastore' or
+                        attr_name.endswith('__datastore') or
+                        attr_name.endswith('__app')):
+                        # Share the reference (don't copy!) to prevent memory leaks
+                        setattr(new_obj, attr_name, attr_value)
+                    # Skip cache attributes - let them regenerate on demand
+                    elif 'cache' in attr_name.lower():
+                        pass  # Don't copy caches
+                    # Copy regular instance attributes
+                    elif not callable(attr_value):
+                        setattr(new_obj, attr_name, attr_value)
+                except AttributeError:
+                    pass  # Attribute doesn't exist in this instance
+
+        return new_obj
+
+    def __getstate__(self):
+        """
+        Custom pickle serialization for all watch_base subclasses.
+
+        Excludes large reference objects (like __datastore) from serialization.
+        """
+        # Get the dict data
+        state = dict(self)
+
+        # Collect instance attributes (excluding methods and large references)
+        instance_attrs = {}
+        for attr_name in dir(self):
+            if attr_name.startswith('_') and not attr_name.startswith('__'):
+                try:
+                    attr_value = getattr(self, attr_name)
+                    # Exclude large reference objects and caches from serialization
+                    if not (attr_name == '_datastore' or
+                           attr_name.endswith('__datastore') or
+                           attr_name.endswith('__app') or
+                           'cache' in attr_name.lower() or
+                           callable(attr_value)):
+                        instance_attrs[attr_name] = attr_value
+                except AttributeError:
+                    pass
+
+        if instance_attrs:
+            state['__instance_metadata__'] = instance_attrs
+
+        return state
+
+    def __setstate__(self, state):
+        """
+        Custom pickle deserialization for all watch_base subclasses.
+
+        WARNING: Large reference objects (like __datastore) are NOT restored!
+        Caller must restore these references after unpickling if needed.
+        """
+        # Extract metadata
+        metadata = state.pop('__instance_metadata__', {})
+
+        # Restore dict data
+        self.update(state)
+
+        # Restore instance attributes
+        for attr_name, attr_value in metadata.items():
+            setattr(self, attr_name, attr_value)
+
+    @property
+    def data_dir(self):
+        """
+        The base directory for this watch/tag data (property, computed from UUID).
+
+        Common property for both Watch and Tag objects.
+        Returns path like: /datastore/{uuid}/
+        """
+        return os.path.join(self._datastore_path, self['uuid']) if self._datastore_path else None
+
+    def ensure_data_dir_exists(self):
+        """
+        Create the data directory if it doesn't exist.
+
+        Common method for both Watch and Tag objects.
+        """
+        from loguru import logger
+        if not os.path.isdir(self.data_dir):
+            logger.debug(f"> Creating data dir {self.data_dir}")
+            os.mkdir(self.data_dir)
+
+    def get_global_setting(self, *path):
+        """
+        Get a setting from the global datastore configuration.
+
+        Args:
+            *path: Path to the setting (e.g., 'application', 'history_snapshot_max_length')
+
+        Returns:
+            The setting value, or None if not found
+
+        Example:
+            maxlen = self.get_global_setting('application', 'history_snapshot_max_length')
+        """
+        if not self._datastore:
+            return None
+
+        try:
+            value = self._datastore['settings']
+            for key in path:
+                value = value[key]
+            return value
+        except (KeyError, TypeError):
+            return None
+
+    def _get_commit_data(self):
+        """
+        Prepare data for commit (can be overridden by subclasses).
+
+        Returns:
+            dict: Data to serialize (filtered as needed by subclass)
+        """
+        import copy
+
+        # Acquire datastore lock to prevent concurrent modifications during copy
+        lock = self._datastore.lock if self._datastore and hasattr(self._datastore, 'lock') else None
+
+        if lock:
+            with lock:
+                snapshot = dict(self)
+        else:
+            snapshot = dict(self)
+
+        # Deep copy snapshot (slower, but done outside lock to minimize contention)
+        # Subclasses can override to filter keys (e.g., Watch excludes processor_config_*)
+        return {k: copy.deepcopy(v) for k, v in snapshot.items()}
+
+    def _save_to_disk(self, data_dict, uuid):
+        """
+        Save data to disk (must be implemented by subclasses).
+
+        Args:
+            data_dict: Dictionary to save
+            uuid: UUID for logging
+
+        Raises:
+            NotImplementedError: If subclass doesn't implement
+        """
+        raise NotImplementedError("Subclass must implement _save_to_disk()")
+
+    def commit(self):
+        """
+        Save this watch/tag immediately to disk using atomic write.
+
+        Common commit logic for Watch and Tag objects.
+        Subclasses override _get_commit_data() and _save_to_disk() for specifics.
+
+        Fire-and-forget: Logs errors but does not raise exceptions.
+        Data remains in memory even if save fails, so next commit will retry.
+        """
+        from loguru import logger
+
+        if not self.data_dir:
+            entity_type = self.__class__.__name__
+            logger.error(f"Cannot commit {entity_type} {self.get('uuid')} without datastore_path")
+            return
+
+        uuid = self.get('uuid')
+        if not uuid:
+            entity_type = self.__class__.__name__
+            logger.error(f"Cannot commit {entity_type} without UUID")
+            return
+
+        # Get data from subclass (may filter keys)
+        try:
+            data_dict = self._get_commit_data()
+        except Exception as e:
+            logger.error(f"Failed to prepare commit data for {uuid}: {e}")
+            return
+
+        # Save to disk via subclass implementation
+        try:
+            # Determine entity type from module name (Watch.py -> watch, Tag.py -> tag)
+            entity_type = _determine_entity_type(self.__class__)
+            filename = f"{entity_type}.json"
+            self._save_to_disk(data_dict, uuid)
+            logger.debug(f"Committed {entity_type} {uuid} to {uuid}/{filename}")
+        except Exception as e:
+            logger.error(f"Failed to commit {uuid}: {e}")

changedetectionio/model/persistence.py

@@ -0,0 +1,84 @@
+"""
+Entity persistence mixin for Watch and Tag models.
+
+Provides file-based persistence using atomic writes.
+"""
+
+import functools
+import inspect
+
+
+@functools.lru_cache(maxsize=None)
+def _determine_entity_type(cls):
+    """
+    Determine entity type from class hierarchy (cached at class level).
+
+    Args:
+        cls: The class to inspect
+
+    Returns:
+        str: Entity type ('watch', 'tag', etc.)
+
+    Raises:
+        ValueError: If entity type cannot be determined
+    """
+    for base_class in inspect.getmro(cls):
+        module_name = base_class.__module__
+        if module_name.startswith('changedetectionio.model.'):
+            # Get last part after dot: "changedetectionio.model.Watch" -> "watch"
+            return module_name.split('.')[-1].lower()
+
+    raise ValueError(
+        f"Cannot determine entity type for {cls.__module__}.{cls.__name__}. "
+        f"Entity must inherit from a class in changedetectionio.model (Watch or Tag)."
+    )
+
+
+class EntityPersistenceMixin:
+    """
+    Mixin providing file persistence for watch_base subclasses (Watch, Tag, etc.).
+
+    This mixin provides the _save_to_disk() method required by watch_base.commit().
+    It automatically determines the correct filename and size limits based on class hierarchy.
+
+    Usage:
+        class model(EntityPersistenceMixin, watch_base):  # in Watch.py
+            pass
+
+        class model(EntityPersistenceMixin, watch_base):  # in Tag.py
+            pass
+    """
+
+    def _save_to_disk(self, data_dict, uuid):
+        """
+        Save entity to disk using atomic write.
+
+        Implements the abstract method required by watch_base.commit().
+        Automatically determines filename and size limits from class hierarchy.
+
+        Args:
+            data_dict: Dictionary to save
+            uuid: UUID for logging
+
+        Raises:
+            ValueError: If entity type cannot be determined from class hierarchy
+        """
+        # Import here to avoid circular dependency
+        from changedetectionio.store.file_saving_datastore import save_entity_atomic
+
+        # Determine entity type (cached at class level, not instance level)
+        entity_type = _determine_entity_type(self.__class__)
+
+        # Set filename and size limits based on entity type
+        filename = f'{entity_type}.json'
+        max_size_mb = 10 if entity_type == 'watch' else 1
+
+        # Save using generic function
+        save_entity_atomic(
+            self.data_dir,
+            uuid,
+            data_dict,
+            filename=filename,
+            entity_type=entity_type,
+            max_size_mb=max_size_mb
+        )

changedetectionio/pluggy_interface.py

@@ -105,6 +105,30 @@ class ChangeDetectionSpec:
         """
         pass
 
+    @hookspec
+    def register_processor(self):
+        """Register an external processor plugin.
+
+        External packages can implement this hook to register custom processors
+        that will be discovered alongside built-in processors.
+
+        Returns:
+            dict or None: Dictionary with processor information:
+                {
+                    'processor_name': str,      # Machine name (e.g., 'osint_recon')
+                    'processor_module': module, # Module containing processor.py
+                    'processor_class': class,   # The perform_site_check class
+                    'metadata': {               # Optional metadata
+                        'name': str,            # Display name
+                        'description': str,     # Description
+                        'processor_weight': int,# Sort weight (lower = higher priority)
+                        'list_badge_text': str, # Badge text for UI
+                    }
+                }
+                Return None if this plugin doesn't provide a processor
+        """
+        pass
+
 
 # Set up Plugin Manager
 plugin_manager = pluggy.PluginManager(PLUGIN_NAMESPACE)

changedetectionio/processors/__init__.py

@@ -17,9 +17,11 @@ def find_sub_packages(package_name):
     return [name for _, name, is_pkg in pkgutil.iter_modules(package.__path__) if is_pkg]
 
 
+@lru_cache(maxsize=1)
 def find_processors():
     """
     Find all subclasses of DifferenceDetectionProcessor in the specified package.
+    Results are cached to avoid repeated discovery.
 
     :param package_name: The name of the package to scan for processor modules.
     :return: A list of (module, class) tuples.
@@ -46,6 +48,23 @@ def find_processors():
         except (ModuleNotFoundError, ImportError) as e:
             logger.warning(f"Failed to import module {module_name}: {e} (find_processors())")
 
+    # Discover plugin processors via pluggy
+    try:
+        from changedetectionio.pluggy_interface import plugin_manager
+        plugin_results = plugin_manager.hook.register_processor()
+
+        for result in plugin_results:
+            if result and isinstance(result, dict):
+                processor_module = result.get('processor_module')
+                processor_name = result.get('processor_name')
+
+                if processor_module and processor_name:
+                    processors.append((processor_module, processor_name))
+                    plugin_path = getattr(processor_module, '__file__', 'unknown location')
+                    logger.info(f"Registered plugin processor: {processor_name} from {plugin_path}")
+    except Exception as e:
+        logger.warning(f"Error loading plugin processors: {e}")
+
     return processors
 
 
@@ -97,54 +116,137 @@ def find_processor_module(processor_name):
     return None
 
 
+def get_processor_module(processor_name):
+    """
+    Get the actual processor module (with perform_site_check class) by name.
+    Works for both built-in and plugin processors.
+
+    Args:
+        processor_name: Processor machine name (e.g., 'text_json_diff', 'osint_recon')
+
+    Returns:
+        module: The processor module containing perform_site_check, or None if not found
+    """
+    processor_classes = find_processors()
+    processor_tuple = next((tpl for tpl in processor_classes if tpl[1] == processor_name), None)
+
+    if processor_tuple:
+        # Return the actual processor module (first element of tuple)
+        return processor_tuple[0]
+
+    return None
+
+
+def get_processor_submodule(processor_name, submodule_name):
+    """
+    Get an optional submodule from a processor (e.g., 'difference', 'extract', 'preview').
+    Works for both built-in and plugin processors.
+
+    Args:
+        processor_name: Processor machine name (e.g., 'text_json_diff', 'osint_recon')
+        submodule_name: Name of the submodule (e.g., 'difference', 'extract', 'preview')
+
+    Returns:
+        module: The submodule if it exists, or None if not found
+    """
+    processor_classes = find_processors()
+    processor_tuple = next((tpl for tpl in processor_classes if tpl[1] == processor_name), None)
+
+    if not processor_tuple:
+        return None
+
+    processor_module = processor_tuple[0]
+    parent_module = get_parent_module(processor_module)
+
+    if not parent_module:
+        return None
+
+    # Try to import the submodule
+    try:
+        # For built-in processors: changedetectionio.processors.text_json_diff.difference
+        # For plugin processors: changedetectionio_osint.difference
+        parent_module_name = parent_module.__name__
+        submodule_full_name = f"{parent_module_name}.{submodule_name}"
+        return importlib.import_module(submodule_full_name)
+    except (ModuleNotFoundError, ImportError):
+        return None
+
+
+@lru_cache(maxsize=1)
+def get_plugin_processor_metadata():
+    """Get metadata from plugin processors."""
+    metadata = {}
+    try:
+        from changedetectionio.pluggy_interface import plugin_manager
+        plugin_results = plugin_manager.hook.register_processor()
+
+        for result in plugin_results:
+            if result and isinstance(result, dict):
+                processor_name = result.get('processor_name')
+                meta = result.get('metadata', {})
+                if processor_name:
+                    metadata[processor_name] = meta
+    except Exception as e:
+        logger.warning(f"Error getting plugin processor metadata: {e}")
+    return metadata
+
+
 def available_processors():
     """
     Get a list of processors by name and description for the UI elements.
-    Can be filtered via ALLOWED_PROCESSORS environment variable (comma-separated list).
+    Can be filtered via DISABLED_PROCESSORS environment variable (comma-separated list).
     :return: A list :)
     """
 
     processor_classes = find_processors()
 
-    # Check if ALLOWED_PROCESSORS env var is set
-    # For now we disable it, need to make a deploy with lots of new code and this will be an overload
-    allowed_processors_env = os.getenv('ALLOWED_PROCESSORS', 'text_json_diff, restock_diff').strip()
-    allowed_processors = None
-    if allowed_processors_env:
+    # Check if DISABLED_PROCESSORS env var is set
+    disabled_processors_env = os.getenv('DISABLED_PROCESSORS', 'image_ssim_diff').strip()
+    disabled_processors = []
+    if disabled_processors_env:
         # Parse comma-separated list and strip whitespace
-        allowed_processors = [p.strip() for p in allowed_processors_env.split(',') if p.strip()]
-        logger.info(f"ALLOWED_PROCESSORS set, filtering to: {allowed_processors}")
+        disabled_processors = [p.strip() for p in disabled_processors_env.split(',') if p.strip()]
+        logger.info(f"DISABLED_PROCESSORS set, disabling: {disabled_processors}")
 
     available = []
+    plugin_metadata = get_plugin_processor_metadata()
+
     for module, sub_package_name in processor_classes:
-        # Filter by allowed processors if set
-        if allowed_processors and sub_package_name not in allowed_processors:
-            logger.debug(f"Skipping processor '{sub_package_name}' (not in ALLOWED_PROCESSORS)")
+        # Skip disabled processors
+        if sub_package_name in disabled_processors:
+            logger.debug(f"Skipping processor '{sub_package_name}' (in DISABLED_PROCESSORS)")
             continue
 
-        # Try to get the 'name' attribute from the processor module first
-        if hasattr(module, 'name'):
-            description = gettext(module.name)
+        # Check if this is a plugin processor
+        if sub_package_name in plugin_metadata:
+            meta = plugin_metadata[sub_package_name]
+            description = gettext(meta.get('name', sub_package_name))
+            # Plugin processors start from weight 10 to separate them from built-in processors
+            weight = 100 + meta.get('processor_weight', 0)
         else:
-            # Fall back to processor_description from parent module's __init__.py
-            parent_module = get_parent_module(module)
-            if parent_module and hasattr(parent_module, 'processor_description'):
-                description = gettext(parent_module.processor_description)
+            # Try to get the 'name' attribute from the processor module first
+            if hasattr(module, 'name'):
+                description = gettext(module.name)
             else:
-                # Final fallback to a readable name
-                description = sub_package_name.replace('_', ' ').title()
+                # Fall back to processor_description from parent module's __init__.py
+                parent_module = get_parent_module(module)
+                if parent_module and hasattr(parent_module, 'processor_description'):
+                    description = gettext(parent_module.processor_description)
+                else:
+                    # Final fallback to a readable name
+                    description = sub_package_name.replace('_', ' ').title()
 
-        # Get weight for sorting (lower weight = higher in list)
-        weight = 0  # Default weight for processors without explicit weight
+            # Get weight for sorting (lower weight = higher in list)
+            weight = 0  # Default weight for processors without explicit weight
 
-        # Check processor module itself first
-        if hasattr(module, 'processor_weight'):
-            weight = module.processor_weight
-        else:
-            # Fall back to parent module (package __init__.py)
-            parent_module = get_parent_module(module)
-            if parent_module and hasattr(parent_module, 'processor_weight'):
-                weight = parent_module.processor_weight
+            # Check processor module itself first
+            if hasattr(module, 'processor_weight'):
+                weight = module.processor_weight
+            else:
+                # Fall back to parent module (package __init__.py)
+                parent_module = get_parent_module(module)
+                if parent_module and hasattr(parent_module, 'processor_weight'):
+                    weight = parent_module.processor_weight
 
         available.append((sub_package_name, description, weight))
 
@@ -155,6 +257,20 @@ def available_processors():
     return [(name, desc) for name, desc, weight in available]
 
 
+def get_default_processor():
+    """
+    Get the default processor to use when none is specified.
+    Returns the first available processor based on weight (lowest weight = highest priority).
+    This ensures forms auto-select a valid processor even when DISABLED_PROCESSORS filters the list.
+
+    :return: The processor name string (e.g., 'text_json_diff')
+    """
+    available = available_processors()
+    if available:
+        return available[0][0]  # Return the processor name from first tuple
+    return 'text_json_diff'  # Fallback if somehow no processors are available
+
+
 def get_processor_badge_texts():
     """
     Get a dictionary mapping processor names to their list_badge_text values.
@@ -279,3 +395,76 @@ def get_processor_badge_css():
 
     return '\n\n'.join(css_rules)
 
+
+def save_processor_config(datastore, watch_uuid, config_data):
+    """
+    Save processor-specific configuration to JSON file.
+
+    This is a shared helper function used by both the UI edit form and API endpoints
+    to consistently handle processor configuration storage.
+
+    Args:
+        datastore: The application datastore instance
+        watch_uuid: UUID of the watch
+        config_data: Dictionary of configuration data to save (with processor_config_* prefix removed)
+
+    Returns:
+        bool: True if saved successfully, False otherwise
+    """
+    if not config_data:
+        return True
+
+    try:
+        from changedetectionio.processors.base import difference_detection_processor
+
+        # Get processor name from watch
+        watch = datastore.data['watching'].get(watch_uuid)
+        if not watch:
+            logger.error(f"Cannot save processor config: watch {watch_uuid} not found")
+            return False
+
+        processor_name = watch.get('processor', 'text_json_diff')
+
+        # Create a processor instance to access config methods
+        processor_instance = difference_detection_processor(datastore, watch_uuid)
+
+        # Use processor name as filename so each processor keeps its own config
+        config_filename = f'{processor_name}.json'
+        processor_instance.update_extra_watch_config(config_filename, config_data)
+
+        logger.debug(f"Saved processor config to {config_filename}: {config_data}")
+        return True
+
+    except Exception as e:
+        logger.error(f"Failed to save processor config: {e}")
+        return False
+
+
+def extract_processor_config_from_form_data(form_data):
+    """
+    Extract processor_config_* fields from form data and return separate dicts.
+
+    This is a shared helper function used by both the UI edit form and API endpoints
+    to consistently handle processor configuration extraction.
+
+    IMPORTANT: This function modifies form_data in-place by removing processor_config_* fields.
+
+    Args:
+        form_data: Dictionary of form data (will be modified in-place)
+
+    Returns:
+        dict: Dictionary of processor config data (with processor_config_* prefix removed)
+    """
+    processor_config_data = {}
+
+    # Use list() to create a copy of keys since we're modifying the dict
+    for field_name in list(form_data.keys()):
+        if field_name.startswith('processor_config_'):
+            config_key = field_name.replace('processor_config_', '')
+            # Save all values (including empty strings) to allow explicit clearing of settings
+            processor_config_data[config_key] = form_data[field_name]
+            # Remove from form_data to prevent it from reaching datastore
+            del form_data[field_name]
+
+    return processor_config_data
+

changedetectionio/processors/base.py

@@ -23,7 +23,14 @@ class difference_detection_processor():
     def __init__(self, datastore, watch_uuid):
         self.datastore = datastore
         self.watch_uuid = watch_uuid
+
+        # Create a stable snapshot of the watch for processing
+        # Why deepcopy?
+        # 1. Prevents "dict changed during iteration" errors if watch is modified during processing
+        # 2. Preserves Watch object with properties (.link, .is_pdf, etc.) - can't use dict()
+        # 3. Safe now: Watch.__deepcopy__() shares datastore ref (no memory leak) but copies dict data
         self.watch = deepcopy(self.datastore.data['watching'].get(watch_uuid))
+
         # Generic fetcher that should be extended (requests, playwright etc)
         self.fetcher = Fetcher()
 
@@ -186,12 +193,12 @@ class difference_detection_processor():
         import os
 
         watch = self.datastore.data['watching'].get(self.watch_uuid)
-        watch_data_dir = watch.watch_data_dir
+        data_dir = watch.data_dir
 
-        if not watch_data_dir:
+        if not data_dir:
             return {}
 
-        filepath = os.path.join(watch_data_dir, filename)
+        filepath = os.path.join(data_dir, filename)
 
         if not os.path.isfile(filepath):
             return {}
@@ -216,16 +223,16 @@ class difference_detection_processor():
         import os
 
         watch = self.datastore.data['watching'].get(self.watch_uuid)
-        watch_data_dir = watch.watch_data_dir
+        data_dir = watch.data_dir
 
-        if not watch_data_dir:
-            logger.warning(f"Cannot save extra watch config {filename}: no watch_data_dir")
+        if not data_dir:
+            logger.warning(f"Cannot save extra watch config {filename}: no data_dir")
             return
 
         # Ensure directory exists
         watch.ensure_data_dir_exists()
 
-        filepath = os.path.join(watch_data_dir, filename)
+        filepath = os.path.join(data_dir, filename)
 
         try:
             # If merge is enabled, read existing data first

changedetectionio/processors/image_ssim_diff/__init__.py

@@ -12,6 +12,13 @@ processor_description = "Visual/Screenshot change detection (Fast)"
 processor_name = "image_ssim_diff"
 processor_weight = 2  # Lower weight = appears at top, heavier weight = appears lower (bottom)
 
+# Processor capabilities
+supports_visual_selector = True
+supports_browser_steps = True
+supports_text_filters_and_triggers = False
+supports_text_filters_and_triggers_elements = False
+supports_request_type = True
+
 PROCESSOR_CONFIG_NAME = f"{Path(__file__).parent.name}.json"
 
 # Subprocess timeout settings

changedetectionio/processors/image_ssim_diff/difference.py

@@ -414,7 +414,7 @@ def render(watch, datastore, request, url_for, render_template, flash, redirect)
 
     # Load historical data if available (for charts/visualization)
     comparison_data = {}
-    comparison_config_path = os.path.join(watch.watch_data_dir, "visual_comparison_data.json")
+    comparison_config_path = os.path.join(watch.data_dir, "visual_comparison_data.json")
     if os.path.isfile(comparison_config_path):
         try:
             with open(comparison_config_path, 'r') as f:

changedetectionio/processors/image_ssim_diff/edit_hook.py

@@ -90,7 +90,7 @@ def on_config_save(watch, processor_config, datastore):
             processor_config['auto_track_region'] = False
 
             # Remove old template file if exists
-            template_path = os.path.join(watch.watch_data_dir, CROPPED_IMAGE_TEMPLATE_FILENAME)
+            template_path = os.path.join(watch.data_dir, CROPPED_IMAGE_TEMPLATE_FILENAME)
             if os.path.exists(template_path):
                 os.remove(template_path)
                 logger.debug(f"Removed old template file: {template_path}")

changedetectionio/processors/restock_diff/__init__.py

@@ -4,6 +4,13 @@ from changedetectionio.model.Watch import model as BaseWatch
 from typing import Union
 import re
 
+# Processor capabilities
+supports_visual_selector = True
+supports_browser_steps = True
+supports_text_filters_and_triggers = True
+supports_text_filters_and_triggers_elements = True
+supports_request_type = True
+
 class Restock(dict):
 
     def parse_currency(self, raw_value: str) -> Union[float, None]:

changedetectionio/processors/restock_diff/processor.py

@@ -193,18 +193,17 @@ class perform_site_check(difference_detection_processor):
 
 
         itemprop_availability = {}
+        multiple_prices_found = False
 
         # Try built-in extraction first, this will scan metadata in the HTML
         try:
             itemprop_availability = get_itemprop_availability(self.fetcher.content)
         except MoreThanOnePriceFound as e:
-            # Add the real data
-            raise ProcessorException(message="Cannot run, more than one price detected, this plugin is only for product pages with ONE product, try the content-change detection mode.",
-                                     url=watch.get('url'),
-                                     status_code=self.fetcher.get_last_status_code(),
-                                     screenshot=self.fetcher.screenshot,
-                                     xpath_data=self.fetcher.xpath_data
-                                     )
+            # Don't raise immediately - let plugins try to handle this case
+            # Plugins might be able to determine which price is correct
+            logger.warning(f"Built-in detection found multiple prices on {watch.get('url')}, will try plugin override")
+            multiple_prices_found = True
+            itemprop_availability = {}
 
         # If built-in extraction didn't get both price AND availability, try plugin override
         # Only check plugin if this watch is using a fetcher that might provide better data
@@ -216,9 +215,21 @@ class perform_site_check(difference_detection_processor):
             from changedetectionio.pluggy_interface import get_itemprop_availability_from_plugin
             fetcher_name = watch.get('fetch_backend', 'html_requests')
 
-            # Only try plugin override if not using system default (which might be anything)
-            if fetcher_name and fetcher_name != 'system':
-                logger.debug("Calling extra plugins for getting item price/availability")
+            # Resolve 'system' to the actual fetcher being used
+            # This allows plugins to work even when watch uses "system settings default"
+            if fetcher_name == 'system':
+                # Get the actual fetcher that was used (from self.fetcher)
+                # Fetcher class name gives us the actual backend (e.g., 'html_requests', 'html_webdriver')
+                actual_fetcher = type(self.fetcher).__name__
+                if 'html_requests' in actual_fetcher.lower():
+                    fetcher_name = 'html_requests'
+                elif 'webdriver' in actual_fetcher.lower() or 'playwright' in actual_fetcher.lower():
+                    fetcher_name = 'html_webdriver'
+                logger.debug(f"Resolved 'system' fetcher to actual fetcher: {fetcher_name}")
+
+            # Try plugin override - plugins can decide if they support this fetcher
+            if fetcher_name:
+                logger.debug(f"Calling extra plugins for getting item price/availability (fetcher: {fetcher_name})")
                 plugin_availability = get_itemprop_availability_from_plugin(self.fetcher.content, fetcher_name, self.fetcher, watch.link)
 
                 if plugin_availability:
@@ -233,6 +244,16 @@ class perform_site_check(difference_detection_processor):
                 if not plugin_availability:
                     logger.debug("No item price/availability from plugins")
 
+        # If we had multiple prices and plugins also failed, NOW raise the exception
+        if multiple_prices_found and not itemprop_availability.get('price'):
+            raise ProcessorException(
+                message="Cannot run, more than one price detected, this plugin is only for product pages with ONE product, try the content-change detection mode.",
+                url=watch.get('url'),
+                status_code=self.fetcher.get_last_status_code(),
+                screenshot=self.fetcher.screenshot,
+                xpath_data=self.fetcher.xpath_data
+            )
+
         # Something valid in get_itemprop_availability() by scraping metadata ?
         if itemprop_availability.get('price') or itemprop_availability.get('availability'):
             # Store for other usage

changedetectionio/processors/text_json_diff/__init__.py

@@ -1,6 +1,12 @@
-
 from loguru import logger
 
+# Processor capabilities
+supports_visual_selector = True
+supports_browser_steps = True
+supports_text_filters_and_triggers = True
+supports_text_filters_and_triggers_elements = True
+supports_request_type = True
+
 
 
 def _task(watch, update_handler):
@@ -49,7 +55,7 @@ def prepare_filter_prevew(datastore, watch_uuid, form_data):
 
     tmp_watch = deepcopy(datastore.data['watching'].get(watch_uuid))
 
-    if tmp_watch and tmp_watch.history and os.path.isdir(tmp_watch.watch_data_dir):
+    if tmp_watch and tmp_watch.history and os.path.isdir(tmp_watch.data_dir):
         # Splice in the temporary stuff from the form
         form = forms.processor_text_json_diff_form(formdata=form_data if request.method == 'POST' else None,
                                                    data=form_data
@@ -58,11 +64,11 @@ def prepare_filter_prevew(datastore, watch_uuid, form_data):
         # Only update vars that came in via the AJAX post
         p = {k: v for k, v in form.data.items() if k in form_data.keys()}
         tmp_watch.update(p)
-        blank_watch_no_filters = watch_model()
+        blank_watch_no_filters = watch_model(datastore_path=datastore.datastore_path, __datastore=datastore.data)
         blank_watch_no_filters['url'] = tmp_watch.get('url')
 
         latest_filename = next(reversed(tmp_watch.history))
-        html_fname = os.path.join(tmp_watch.watch_data_dir, f"{latest_filename}.html.br")
+        html_fname = os.path.join(tmp_watch.data_dir, f"{latest_filename}.html.br")
         with open(html_fname, 'rb') as f:
             decompressed_data = brotli.decompress(f.read()).decode('utf-8') if html_fname.endswith('.br') else f.read().decode('utf-8')
 

changedetectionio/queue_handlers.py

@@ -5,51 +5,57 @@ import heapq
 import queue
 import threading
 
-try:
-    import janus
-except ImportError:
-    logger.critical(f"CRITICAL: janus library is required. Install with: pip install janus")
-    raise
+# Janus is no longer required - we use pure threading.Queue for multi-loop support
+# try:
+#     import janus
+# except ImportError:
+#     pass  # Not needed anymore
 
 
 class RecheckPriorityQueue:
     """
-    Ultra-reliable priority queue using janus for async/sync bridging.
-    
-    CRITICAL DESIGN NOTE: Both sync_q and async_q are required because:
-    - sync_q: Used by Flask routes, ticker threads, and other synchronous code
-    - async_q: Used by async workers (the actual fetchers/processors) and coroutines
-    
-    DO NOT REMOVE EITHER INTERFACE - they bridge different execution contexts:
-    - Synchronous code (Flask, threads) cannot use async methods without blocking
-    - Async code cannot use sync methods without blocking the event loop
-    - janus provides the only safe bridge between these two worlds
-    
-    Attempting to unify to async-only would require:
-    - Converting all Flask routes to async (major breaking change)
-    - Using asyncio.run() in sync contexts (causes deadlocks)
-    - Thread-pool wrapping (adds complexity and overhead)
-    
-    Minimal implementation focused on reliability:
-    - Pure janus for sync/async bridge
-    - Thread-safe priority ordering  
-    - Bulletproof error handling with critical logging
+    Thread-safe priority queue supporting multiple async event loops.
+
+    ARCHITECTURE:
+    - Multiple async workers, each with its own event loop in its own thread
+    - Hybrid sync/async design for maximum scalability
+    - Sync interface for ticker thread (threading.Queue)
+    - Async interface for workers (asyncio.Event - NO executor threads!)
+
+    SCALABILITY:
+    - Scales to 100-200+ workers without executor thread exhaustion
+    - Async workers wait on asyncio.Event (pure coroutines, no threads)
+    - Sync callers use threading.Queue (backward compatible)
+
+    WHY NOT JANUS:
+    - Janus binds to ONE event loop at creation time
+    - Our architecture has 15+ workers, each with separate event loops
+    - Workers in different threads/loops cannot share janus async interface
+
+    WHY NOT RUN_IN_EXECUTOR:
+    - With 200 workers, run_in_executor() would block 200 threads
+    - Exhausts ThreadPoolExecutor, starves Flask HTTP handlers
+    - Pure async approach uses 0 threads while waiting
     """
-    
+
     def __init__(self, maxsize: int = 0):
         try:
-            self._janus_queue = janus.Queue(maxsize=maxsize)
-            # BOTH interfaces required - see class docstring for why
-            self.sync_q = self._janus_queue.sync_q   # Flask routes, ticker thread
-            self.async_q = self._janus_queue.async_q # Async workers
-            
+            import asyncio
+
+            # Sync interface: threading.Queue for ticker thread and Flask routes
+            self._notification_queue = queue.Queue(maxsize=maxsize if maxsize > 0 else 0)
+
             # Priority storage - thread-safe
             self._priority_items = []
             self._lock = threading.RLock()
-            
+
+            # No event signaling needed - pure polling approach
+            # Workers check queue every 50ms (latency acceptable: 0-500ms)
+            # Scales to 1000+ workers: each sleeping worker = ~4KB coroutine, not thread
+
             # Signals for UI updates
             self.queue_length_signal = signal('queue_length')
-            
+
             logger.debug("RecheckPriorityQueue initialized successfully")
         except Exception as e:
             logger.critical(f"CRITICAL: Failed to initialize RecheckPriorityQueue: {str(e)}")
@@ -58,38 +64,48 @@ class RecheckPriorityQueue:
     # SYNC INTERFACE (for ticker thread)
     def put(self, item, block: bool = True, timeout: Optional[float] = None):
         """Thread-safe sync put with priority ordering"""
+        logger.trace(f"RecheckQueue.put() called for item: {self._get_item_uuid(item)}, block={block}, timeout={timeout}")
         try:
-            # Add to priority storage
+            # CRITICAL: Add to both priority storage AND notification queue atomically
+            # to prevent desynchronization where item exists but no notification
             with self._lock:
                 heapq.heappush(self._priority_items, item)
-            
-            # Notify via janus sync queue
-            self.sync_q.put(True, block=block, timeout=timeout)
-            
-            # Emit signals
-            self._emit_put_signals(item)
-            
+
+                # Add notification - use blocking with timeout for safety
+                # Notification queue is unlimited size, so should never block in practice
+                # but timeout ensures we detect any unexpected issues (deadlock, etc)
+                try:
+                    self._notification_queue.put(True, block=True, timeout=5.0)
+                except Exception as notif_e:
+                    # Notification failed - MUST remove from priority_items to keep in sync
+                    # This prevents "Priority queue inconsistency" errors in get()
+                    logger.critical(f"CRITICAL: Notification queue put failed, removing from priority_items: {notif_e}")
+                    self._priority_items.remove(item)
+                    heapq.heapify(self._priority_items)
+                    raise  # Re-raise to be caught by outer exception handler
+
+            # Signal emission after successful queue - log but don't fail the operation
+            # Item is already safely queued, so signal failure shouldn't affect queue state
+            try:
+                self._emit_put_signals(item)
+            except Exception as signal_e:
+                logger.error(f"Failed to emit put signals but item queued successfully: {signal_e}")
+
             logger.trace(f"Successfully queued item: {self._get_item_uuid(item)}")
             return True
-            
+
         except Exception as e:
-            logger.critical(f"CRITICAL: Failed to put item {self._get_item_uuid(item)}: {str(e)}")
-            # Remove from priority storage if janus put failed
-            try:
-                with self._lock:
-                    if item in self._priority_items:
-                        self._priority_items.remove(item)
-                        heapq.heapify(self._priority_items)
-            except Exception as cleanup_e:
-                logger.critical(f"CRITICAL: Failed to cleanup after put failure: {str(e)}")
+            logger.critical(f"CRITICAL: Failed to put item {self._get_item_uuid(item)}: {type(e).__name__}: {str(e)}")
+            # Item should have been cleaned up in the inner try/except if notification failed
             return False
     
     def get(self, block: bool = True, timeout: Optional[float] = None):
         """Thread-safe sync get with priority ordering"""
-        import queue
+        logger.trace(f"RecheckQueue.get() called, block={block}, timeout={timeout}")
+        import queue as queue_module
         try:
-            # Wait for notification
-            self.sync_q.get(block=block, timeout=timeout)
+            # Wait for notification (this doesn't return the actual item, just signals availability)
+            self._notification_queue.get(block=block, timeout=timeout)
 
             # Get highest priority item
             with self._lock:
@@ -98,69 +114,91 @@ class RecheckPriorityQueue:
                     raise Exception("Priority queue inconsistency")
                 item = heapq.heappop(self._priority_items)
 
-            # Emit signals
-            self._emit_get_signals()
+            # Signal emission after successful retrieval - log but don't lose the item
+            # Item is already retrieved, so signal failure shouldn't affect queue state
+            try:
+                self._emit_get_signals()
+            except Exception as signal_e:
+                logger.error(f"Failed to emit get signals but item retrieved successfully: {signal_e}")
 
-            logger.debug(f"Successfully retrieved item: {self._get_item_uuid(item)}")
+            logger.trace(f"RecheckQueue.get() successfully retrieved item: {self._get_item_uuid(item)}")
             return item
 
-        except queue.Empty:
-            # Queue is empty with timeout - expected behavior, re-raise without logging
-            raise
+        except queue_module.Empty:
+            # Queue is empty with timeout - expected behavior
+            logger.trace(f"RecheckQueue.get() timed out - queue is empty (timeout={timeout})")
+            raise  # noqa
         except Exception as e:
             # Re-raise without logging - caller (worker) will handle and log appropriately
+            logger.trace(f"RecheckQueue.get() failed with exception: {type(e).__name__}: {str(e)}")
             raise
     
     # ASYNC INTERFACE (for workers)
-    async def async_put(self, item):
-        """Pure async put with priority ordering"""
+    async def async_put(self, item, executor=None):
+        """Async put with priority ordering - uses thread pool to avoid blocking
+
+        Args:
+            item: Item to add to queue
+            executor: Optional ThreadPoolExecutor. If None, uses default pool.
+        """
+        logger.trace(f"RecheckQueue.async_put() called for item: {self._get_item_uuid(item)}, executor={executor}")
+        import asyncio
         try:
-            # Add to priority storage
-            with self._lock:
-                heapq.heappush(self._priority_items, item)
-            
-            # Notify via janus async queue
-            await self.async_q.put(True)
-            
-            # Emit signals
-            self._emit_put_signals(item)
-            
-            logger.debug(f"Successfully async queued item: {self._get_item_uuid(item)}")
-            return True
-            
+            # Use run_in_executor to call sync put without blocking event loop
+            loop = asyncio.get_event_loop()
+            result = await loop.run_in_executor(
+                executor,  # Use provided executor or default
+                lambda: self.put(item, block=True, timeout=5.0)
+            )
+
+            logger.trace(f"RecheckQueue.async_put() successfully queued item: {self._get_item_uuid(item)}")
+            return result
+
         except Exception as e:
             logger.critical(f"CRITICAL: Failed to async put item {self._get_item_uuid(item)}: {str(e)}")
-            # Remove from priority storage if janus put failed
-            try:
-                with self._lock:
-                    if item in self._priority_items:
-                        self._priority_items.remove(item)
-                        heapq.heapify(self._priority_items)
-            except Exception as cleanup_e:
-                logger.critical(f"CRITICAL: Failed to cleanup after async put failure: {str(e)}")
             return False
-    
-    async def async_get(self):
-        """Pure async get with priority ordering"""
+
+    async def async_get(self, executor=None, timeout=1.0):
+        """
+        Efficient async get using executor for blocking call.
+
+        HYBRID APPROACH: Best of both worlds
+        - Uses run_in_executor for efficient blocking (no polling overhead)
+        - Single timeout (no double-timeout race condition)
+        - Scales well: executor sized to match worker count
+
+        With FETCH_WORKERS=10: 10 threads blocked max (acceptable)
+        With FETCH_WORKERS=200: Need executor with 200+ threads (see worker_pool.py)
+
+        Args:
+            executor: ThreadPoolExecutor (sized to match worker count)
+            timeout: Maximum time to wait in seconds
+
+        Returns:
+            Item from queue
+
+        Raises:
+            queue.Empty: If timeout expires with no item available
+        """
+        logger.trace(f"RecheckQueue.async_get() called, timeout={timeout}")
+        import asyncio
         try:
-            # Wait for notification
-            await self.async_q.get()
+            # Use run_in_executor to call sync get efficiently
+            # No outer asyncio.wait_for wrapper = no double timeout issue!
+            loop = asyncio.get_event_loop()
+            item = await loop.run_in_executor(
+                executor,
+                lambda: self.get(block=True, timeout=timeout)
+            )
 
-            # Get highest priority item
-            with self._lock:
-                if not self._priority_items:
-                    logger.critical(f"CRITICAL: Async queue notification received but no priority items available")
-                    raise Exception("Priority queue inconsistency")
-                item = heapq.heappop(self._priority_items)
-
-            # Emit signals
-            self._emit_get_signals()
-
-            logger.debug(f"Successfully async retrieved item: {self._get_item_uuid(item)}")
+            logger.trace(f"RecheckQueue.async_get() successfully retrieved item: {self._get_item_uuid(item)}")
             return item
 
+        except queue.Empty:
+            logger.trace(f"RecheckQueue.async_get() timed out - queue is empty")
+            raise
         except Exception as e:
-            logger.critical(f"CRITICAL: Failed to async get item from queue: {str(e)}")
+            logger.critical(f"CRITICAL: Failed to async get item from queue: {type(e).__name__}: {str(e)}")
             raise
     
     # UTILITY METHODS
@@ -186,10 +224,35 @@ class RecheckPriorityQueue:
             logger.critical(f"CRITICAL: Failed to get queued UUIDs: {str(e)}")
             return []
 
-    def close(self):
-        """Close the janus queue"""
+    def clear(self):
+        """Clear all items from both priority storage and notification queue"""
         try:
-            self._janus_queue.close()
+            with self._lock:
+                # Clear priority items
+                self._priority_items.clear()
+
+                # Drain all notifications to prevent stale notifications
+                # This is critical for test cleanup to prevent queue desynchronization
+                drained = 0
+                while not self._notification_queue.empty():
+                    try:
+                        self._notification_queue.get_nowait()
+                        drained += 1
+                    except queue.Empty:
+                        break
+
+                if drained > 0:
+                    logger.debug(f"Cleared queue: removed {drained} notifications")
+
+            return True
+        except Exception as e:
+            logger.critical(f"CRITICAL: Failed to clear queue: {str(e)}")
+            return False
+
+    def close(self):
+        """Close the queue"""
+        try:
+            # Nothing to close for threading.Queue
             logger.debug("RecheckPriorityQueue closed successfully")
         except Exception as e:
             logger.critical(f"CRITICAL: Failed to close RecheckPriorityQueue: {str(e)}")
@@ -321,7 +384,7 @@ class RecheckPriorityQueue:
         except Exception:
             pass
         return 'unknown'
-    
+
     def _emit_put_signals(self, item):
         """Emit signals when item is added"""
         try:
@@ -330,14 +393,14 @@ class RecheckPriorityQueue:
                 watch_check_update = signal('watch_check_update')
                 if watch_check_update:
                     watch_check_update.send(watch_uuid=item.item['uuid'])
-            
+
             # Queue length signal
             if self.queue_length_signal:
                 self.queue_length_signal.send(length=self.qsize())
-                
+
         except Exception as e:
             logger.critical(f"CRITICAL: Failed to emit put signals: {str(e)}")
-    
+
     def _emit_get_signals(self):
         """Emit signals when item is removed"""
         try:
@@ -363,12 +426,11 @@ class NotificationQueue:
     
     def __init__(self, maxsize: int = 0, datastore=None):
         try:
-            self._janus_queue = janus.Queue(maxsize=maxsize)
-            # BOTH interfaces required - see class docstring for why
-            self.sync_q = self._janus_queue.sync_q   # Flask routes, threads
-            self.async_q = self._janus_queue.async_q # Async workers
+            # Use pure threading.Queue to avoid event loop binding issues
+            self._notification_queue = queue.Queue(maxsize=maxsize if maxsize > 0 else 0)
             self.notification_event_signal = signal('notification_event')
             self.datastore = datastore  # For checking all_muted setting
+            self._lock = threading.RLock()
             logger.debug("NotificationQueue initialized successfully")
         except Exception as e:
             logger.critical(f"CRITICAL: Failed to initialize NotificationQueue: {str(e)}")
@@ -380,72 +442,97 @@ class NotificationQueue:
     
     def put(self, item: Dict[str, Any], block: bool = True, timeout: Optional[float] = None):
         """Thread-safe sync put with signal emission"""
+        logger.trace(f"NotificationQueue.put() called for item: {item.get('uuid', 'unknown')}, block={block}, timeout={timeout}")
         try:
             # Check if all notifications are muted
             if self.datastore and self.datastore.data['settings']['application'].get('all_muted', False):
                 logger.debug(f"Notification blocked - all notifications are muted: {item.get('uuid', 'unknown')}")
                 return False
 
-            self.sync_q.put(item, block=block, timeout=timeout)
+            with self._lock:
+                self._notification_queue.put(item, block=block, timeout=timeout)
             self._emit_notification_signal(item)
-            logger.debug(f"Successfully queued notification: {item.get('uuid', 'unknown')}")
+            logger.trace(f"NotificationQueue.put() successfully queued notification: {item.get('uuid', 'unknown')}")
             return True
         except Exception as e:
             logger.critical(f"CRITICAL: Failed to put notification {item.get('uuid', 'unknown')}: {str(e)}")
             return False
     
-    async def async_put(self, item: Dict[str, Any]):
-        """Pure async put with signal emission"""
+    async def async_put(self, item: Dict[str, Any], executor=None):
+        """Async put with signal emission - uses thread pool
+
+        Args:
+            item: Notification item to queue
+            executor: Optional ThreadPoolExecutor
+        """
+        logger.trace(f"NotificationQueue.async_put() called for item: {item.get('uuid', 'unknown')}, executor={executor}")
+        import asyncio
         try:
             # Check if all notifications are muted
             if self.datastore and self.datastore.data['settings']['application'].get('all_muted', False):
                 logger.debug(f"Notification blocked - all notifications are muted: {item.get('uuid', 'unknown')}")
                 return False
 
-            await self.async_q.put(item)
-            self._emit_notification_signal(item)
-            logger.debug(f"Successfully async queued notification: {item.get('uuid', 'unknown')}")
+            loop = asyncio.get_event_loop()
+            await loop.run_in_executor(executor, lambda: self.put(item, block=True, timeout=5.0))
+            logger.trace(f"NotificationQueue.async_put() successfully queued notification: {item.get('uuid', 'unknown')}")
             return True
         except Exception as e:
             logger.critical(f"CRITICAL: Failed to async put notification {item.get('uuid', 'unknown')}: {str(e)}")
             return False
-    
+
     def get(self, block: bool = True, timeout: Optional[float] = None):
         """Thread-safe sync get"""
+        logger.trace(f"NotificationQueue.get() called, block={block}, timeout={timeout}")
         try:
-            return self.sync_q.get(block=block, timeout=timeout)
+            with self._lock:
+                item = self._notification_queue.get(block=block, timeout=timeout)
+            logger.trace(f"NotificationQueue.get() retrieved item: {item.get('uuid', 'unknown') if isinstance(item, dict) else 'unknown'}")
+            return item
         except queue.Empty as e:
+            logger.trace(f"NotificationQueue.get() timed out - queue is empty (timeout={timeout})")
             raise e
         except Exception as e:
-            logger.critical(f"CRITICAL: Failed to get notification: {str(e)}")
+            logger.critical(f"CRITICAL: Failed to get notification: {type(e).__name__}: {str(e)}")
             raise e
-    
-    async def async_get(self):
-        """Pure async get"""
+
+    async def async_get(self, executor=None):
+        """Async get - uses thread pool
+
+        Args:
+            executor: Optional ThreadPoolExecutor
+        """
+        logger.trace(f"NotificationQueue.async_get() called, executor={executor}")
+        import asyncio
         try:
-            return await self.async_q.get()
+            loop = asyncio.get_event_loop()
+            item = await loop.run_in_executor(executor, lambda: self.get(block=True, timeout=1.0))
+            logger.trace(f"NotificationQueue.async_get() retrieved item: {item.get('uuid', 'unknown') if isinstance(item, dict) else 'unknown'}")
+            return item
         except queue.Empty as e:
+            logger.trace(f"NotificationQueue.async_get() timed out - queue is empty")
             raise e
         except Exception as e:
-            logger.critical(f"CRITICAL: Failed to async get notification: {str(e)}")
+            logger.critical(f"CRITICAL: Failed to async get notification: {type(e).__name__}: {str(e)}")
             raise e
     
     def qsize(self) -> int:
         """Get current queue size"""
         try:
-            return self.sync_q.qsize()
+            with self._lock:
+                return self._notification_queue.qsize()
         except Exception as e:
             logger.critical(f"CRITICAL: Failed to get notification queue size: {str(e)}")
             return 0
-    
+
     def empty(self) -> bool:
         """Check if queue is empty"""
         return self.qsize() == 0
-    
+
     def close(self):
-        """Close the janus queue"""
+        """Close the queue"""
         try:
-            self._janus_queue.close()
+            # Nothing to close for threading.Queue
             logger.debug("NotificationQueue closed successfully")
         except Exception as e:
             logger.critical(f"CRITICAL: Failed to close NotificationQueue: {str(e)}")

changedetectionio/realtime/events.py

@@ -37,9 +37,9 @@ def register_watch_operation_handlers(socketio, datastore):
                 # Import here to avoid circular imports
                 from changedetectionio.flask_app import update_q
                 from changedetectionio import queuedWatchMetaData
-                from changedetectionio import worker_handler
+                from changedetectionio import worker_pool
                 
-                worker_handler.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
+                worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
                 logger.info(f"Socket.IO: Queued recheck for watch {uuid}")
             else:
                 emit('operation_result', {'success': False, 'error': f'Unknown operation: {op}'})

changedetectionio/realtime/socket_server.py

@@ -145,10 +145,10 @@ def handle_watch_update(socketio, **kwargs):
         # Emit the watch update to all connected clients
         from changedetectionio.flask_app import update_q
         from changedetectionio.flask_app import _jinja2_filter_datetime
-        from changedetectionio import worker_handler
+        from changedetectionio import worker_pool
 
         # Get list of watches that are currently running
-        running_uuids = worker_handler.get_running_uuids()
+        running_uuids = worker_pool.get_running_uuids()
 
         # Get list of watches in the queue (efficient single-lock method)
         queue_list = update_q.get_queued_uuids()
@@ -252,7 +252,7 @@ def init_socketio(app, datastore):
     def event_checkbox_operations(data):
         from changedetectionio.blueprint.ui import _handle_operations
         from changedetectionio import queuedWatchMetaData
-        from changedetectionio import worker_handler
+        from changedetectionio import worker_pool
         from changedetectionio.flask_app import update_q, watch_check_update
         import threading
 
@@ -268,7 +268,7 @@ def init_socketio(app, datastore):
                     uuids=data.get('uuids'),
                     datastore=datastore,
                     extra_data=data.get('extra_data'),
-                    worker_handler=worker_handler,
+                    worker_pool=worker_pool,
                     update_q=update_q,
                     queuedWatchMetaData=queuedWatchMetaData,
                     watch_check_update=watch_check_update,

changedetectionio/run_basic_tests.sh

@@ -10,6 +10,7 @@
 set -e
 
 SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )
+rm tests/logs/* -f
 
 # Since theres no curl installed lets roll with python3
 check_sanity() {
@@ -64,18 +65,21 @@ data_sanity_test
 echo "-------------------- Running rest of tests in parallel -------------------------------"
 
 # REMOVE_REQUESTS_OLD_SCREENSHOTS disabled so that we can write a screenshot and send it in test_notifications.py without a real browser
-REMOVE_REQUESTS_OLD_SCREENSHOTS=false \
+FETCH_WORKERS=2 REMOVE_REQUESTS_OLD_SCREENSHOTS=false \
 pytest tests/test_*.py \
-  -n 30 \
+  -n 8 \
   --dist=load \
   -vvv \
   -s \
   --capture=no \
+  -k "not test_queue_system" \
   --log-cli-level=DEBUG \
   --log-cli-format="%(asctime)s [%(process)d] [%(levelname)s] %(name)s: %(message)s"
 
 echo "---------------------------- DONE parallel test ---------------------------------------"
 
+FETCH_WORKERS=20 pytest -vvv -s tests/test_queue_handler.py
+
 echo "RUNNING WITH BASE_URL SET"
 
 # Now re-run some tests with BASE_URL enabled

changedetectionio/static/favicons/site.webmanifest

@@ -1,19 +1,25 @@
 {
-    "name": "",
-    "short_name": "",
+    "name": "ChangeDetection.io",
+    "short_name": "ChangeDetect",
+    "description": "Self-hosted website change detection and monitoring",
     "icons": [
         {
             "src": "android-chrome-192x192.png",
             "sizes": "192x192",
-            "type": "image/png"
+            "type": "image/png",
+            "purpose": "any maskable"
         },
         {
             "src": "android-chrome-256x256.png",
             "sizes": "256x256",
-            "type": "image/png"
+            "type": "image/png",
+            "purpose": "any maskable"
         }
     ],
-    "theme_color": "#ffffff",
+    "start_url": "/",
+    "theme_color": "#5bbad5",
     "background_color": "#ffffff",
-    "display": "standalone"
+    "display": "standalone",
+    "categories": ["utilities", "productivity"],
+    "orientation": "any"
 }

changedetectionio/static/js/modal.js

@@ -184,7 +184,8 @@ $(document).ready(function() {
         }
         // If it's a button in a form, submit the form
         else if ($element.is('button')) {
-          $element.closest('form').submit();
+          // Use requestSubmit() to include the button's name/value in the form data
+          $element.closest('form')[0].requestSubmit($element[0]);
         }
       }
     };

changedetectionio/static/styles/scss/styles.scss

@@ -222,6 +222,19 @@ code {
   color: var(--color-white);
   background: var(--color-text-watch-tag-list);
   @extend .inline-tag;
+  
+  /* Remove default anchor styling when used as links */
+  text-decoration: none;
+  
+  &:hover {
+    text-decoration: none;
+    opacity: 0.8;
+    cursor: pointer;
+  }
+  
+  &:visited {
+    color: var(--color-white);
+  }
 }
 
 @media (min-width: 768px) {

changedetectionio/store/base.py

@@ -0,0 +1,83 @@
+"""
+Base classes for the datastore.
+
+This module defines the abstract interfaces that all datastore implementations must follow.
+"""
+
+from abc import ABC, abstractmethod
+from threading import Lock
+from loguru import logger
+
+
+class DataStore(ABC):
+    """
+    Abstract base class for all datastore implementations.
+
+    Defines the core interface that all datastores must implement for:
+    - Loading and saving data
+    - Managing watches
+    - Handling settings
+    - Providing data access
+    """
+
+    lock = Lock()
+    datastore_path = None
+
+    @abstractmethod
+    def reload_state(self, datastore_path, include_default_watches, version_tag):
+        """
+        Load data from persistent storage.
+
+        Args:
+            datastore_path: Path to the datastore directory
+            include_default_watches: Whether to create default watches if none exist
+            version_tag: Application version string
+        """
+        pass
+
+    @abstractmethod
+    def add_watch(self, url, **kwargs):
+        """
+        Add a new watch.
+
+        Args:
+            url: URL to watch
+            **kwargs: Additional watch parameters
+
+        Returns:
+            UUID of the created watch
+        """
+        pass
+
+    @abstractmethod
+    def update_watch(self, uuid, update_obj):
+        """
+        Update an existing watch.
+
+        Args:
+            uuid: Watch UUID
+            update_obj: Dictionary of fields to update
+        """
+        pass
+
+    @abstractmethod
+    def delete(self, uuid):
+        """
+        Delete a watch.
+
+        Args:
+            uuid: Watch UUID to delete
+        """
+        pass
+
+    @property
+    @abstractmethod
+    def data(self):
+        """
+        Access to the underlying data structure.
+
+        Returns:
+            Dictionary containing all datastore data
+        """
+        pass
+

changedetectionio/store/file_saving_datastore.py

@@ -0,0 +1,509 @@
+"""
+File-based datastore with individual watch persistence and immediate commits.
+
+This module provides the FileSavingDataStore abstract class that implements:
+- Individual watch.json file persistence
+- Immediate commit-based persistence (watch.commit(), datastore.commit())
+- Atomic file writes safe for NFS/NAS
+"""
+
+import glob
+import json
+import os
+import tempfile
+import time
+from loguru import logger
+
+from .base import DataStore
+from .. import strtobool
+
+# Try to import orjson for faster JSON serialization
+try:
+    import orjson
+    HAS_ORJSON = True
+except ImportError:
+    HAS_ORJSON = False
+
+# Fsync configuration: Force file data to disk for crash safety
+# Default False to match legacy behavior (write-and-rename without fsync)
+# Set to True for mission-critical deployments requiring crash consistency
+FORCE_FSYNC_DATA_IS_CRITICAL = bool(strtobool(os.getenv('FORCE_FSYNC_DATA_IS_CRITICAL', 'False')))
+
+# ============================================================================
+# Helper Functions for Atomic File Operations
+# ============================================================================
+
+def save_json_atomic(file_path, data_dict, label="file", max_size_mb=10):
+    """
+    Save JSON data to disk using atomic write pattern.
+
+    Generic helper for saving any JSON data (settings, watches, etc.) with:
+    - Atomic write (temp file + rename)
+    - Directory fsync for crash consistency (only for new files)
+    - Size validation
+    - Proper error handling
+
+    Thread safety: Caller must hold datastore.lock to prevent concurrent modifications.
+    Multi-process safety: Not supported - run only one app instance per datastore.
+
+    Args:
+        file_path: Full path to target JSON file
+        data_dict: Dictionary to serialize
+        label: Human-readable label for error messages (e.g., "watch", "settings")
+        max_size_mb: Maximum allowed file size in MB
+
+    Raises:
+        ValueError: If serialized data exceeds max_size_mb
+        OSError: If disk is full (ENOSPC) or other I/O error
+    """
+    # Check if file already exists (before we start writing)
+    # Directory fsync only needed for NEW files to persist the filename
+    file_exists = os.path.exists(file_path)
+
+    # Ensure parent directory exists
+    parent_dir = os.path.dirname(file_path)
+    os.makedirs(parent_dir, exist_ok=True)
+
+    # Create temp file in same directory (required for NFS atomicity)
+    fd, temp_path = tempfile.mkstemp(
+        suffix='.tmp',
+        prefix='json-',
+        dir=parent_dir,
+        text=False
+    )
+
+    fd_closed = False
+    try:
+        # Serialize data
+        t0 = time.time()
+        if HAS_ORJSON:
+            data = orjson.dumps(data_dict, option=orjson.OPT_INDENT_2)
+        else:
+            data = json.dumps(data_dict, indent=2, ensure_ascii=False).encode('utf-8')
+        serialize_ms = (time.time() - t0) * 1000
+
+        # Safety check: validate size
+        MAX_SIZE = max_size_mb * 1024 * 1024
+        data_size = len(data)
+        if data_size > MAX_SIZE:
+            raise ValueError(
+                f"{label.capitalize()} data is unexpectedly large: {data_size / 1024 / 1024:.2f}MB "
+                f"(max: {max_size_mb}MB). This indicates a bug or data corruption."
+            )
+
+        # Write to temp file
+        t1 = time.time()
+        os.write(fd, data)
+        write_ms = (time.time() - t1) * 1000
+
+        # Optional fsync: Force file data to disk for crash safety
+        # Only if FORCE_FSYNC_DATA_IS_CRITICAL=True (default: False, matches legacy behavior)
+        t2 = time.time()
+        if FORCE_FSYNC_DATA_IS_CRITICAL:
+            os.fsync(fd)
+        file_fsync_ms = (time.time() - t2) * 1000
+
+        os.close(fd)
+        fd_closed = True
+
+        # Atomic rename
+        t3 = time.time()
+        os.replace(temp_path, file_path)
+        rename_ms = (time.time() - t3) * 1000
+
+        # Sync directory to ensure filename metadata is durable
+        # OPTIMIZATION: Only needed for NEW files. Existing files already have
+        # directory entry persisted, so we only need file fsync for data durability.
+        dir_fsync_ms = 0
+        if not file_exists:
+            try:
+                dir_fd = os.open(parent_dir, os.O_RDONLY)
+                try:
+                    t4 = time.time()
+                    os.fsync(dir_fd)
+                    dir_fsync_ms = (time.time() - t4) * 1000
+                finally:
+                    os.close(dir_fd)
+            except (OSError, AttributeError):
+                # Windows doesn't support fsync on directories
+                pass
+
+        # Log timing breakdown for slow saves
+#        total_ms = serialize_ms + write_ms + file_fsync_ms + rename_ms + dir_fsync_ms
+#        if total_ms:  # Log if save took more than 10ms
+#            file_status = "new" if not file_exists else "update"
+#            logger.trace(
+#                f"Save timing breakdown ({total_ms:.1f}ms total, {file_status}): "
+#                f"serialize={serialize_ms:.1f}ms, write={write_ms:.1f}ms, "
+#                f"file_fsync={file_fsync_ms:.1f}ms, rename={rename_ms:.1f}ms, "
+#                f"dir_fsync={dir_fsync_ms:.1f}ms, using_orjson={HAS_ORJSON}"
+#            )
+
+    except OSError as e:
+        # Cleanup temp file
+        if not fd_closed:
+            try:
+                os.close(fd)
+            except:
+                pass
+        if os.path.exists(temp_path):
+            try:
+                os.unlink(temp_path)
+            except:
+                pass
+
+        # Provide helpful error messages
+        if e.errno == 28:  # ENOSPC
+            raise OSError(f"Disk full: Cannot save {label}") from e
+        elif e.errno == 122:  # EDQUOT
+            raise OSError(f"Disk quota exceeded: Cannot save {label}") from e
+        else:
+            raise OSError(f"I/O error saving {label}: {e}") from e
+
+    except Exception as e:
+        # Cleanup temp file
+        if not fd_closed:
+            try:
+                os.close(fd)
+            except:
+                pass
+        if os.path.exists(temp_path):
+            try:
+                os.unlink(temp_path)
+            except:
+                pass
+        raise e
+
+
+def save_entity_atomic(entity_dir, uuid, entity_dict, filename, entity_type, max_size_mb):
+    """
+    Save an entity (watch/tag) to disk using atomic write pattern.
+
+    Generic function for saving any watch_base subclass (Watch, Tag, etc.).
+
+    Args:
+        entity_dir: Directory for this entity (e.g., /datastore/{uuid})
+        uuid: Entity UUID (for logging)
+        entity_dict: Dictionary representation of the entity
+        filename: JSON filename (e.g., 'watch.json', 'tag.json')
+        entity_type: Type label for logging (e.g., 'watch', 'tag')
+        max_size_mb: Maximum allowed file size in MB
+
+    Raises:
+        ValueError: If serialized data exceeds max_size_mb
+        OSError: If disk is full (ENOSPC) or other I/O error
+    """
+    entity_json = os.path.join(entity_dir, filename)
+    save_json_atomic(entity_json, entity_dict, label=f"{entity_type} {uuid}", max_size_mb=max_size_mb)
+
+
+def save_watch_atomic(watch_dir, uuid, watch_dict):
+    """
+    Save a watch to disk using atomic write pattern.
+
+    Convenience wrapper around save_entity_atomic for watches.
+    Kept for backwards compatibility.
+    """
+    save_entity_atomic(watch_dir, uuid, watch_dict, "watch.json", "watch", max_size_mb=10)
+
+
+
+def load_watch_from_file(watch_json, uuid, rehydrate_entity_func):
+    """
+    Load a watch from its JSON file.
+
+    Args:
+        watch_json: Path to the watch.json file
+        uuid: Watch UUID
+        rehydrate_entity_func: Function to convert dict to Watch object
+
+    Returns:
+        Watch object or None if failed
+    """
+    try:
+        # Check file size before reading
+        file_size = os.path.getsize(watch_json)
+        MAX_WATCH_SIZE = 10 * 1024 * 1024  # 10MB
+        if file_size > MAX_WATCH_SIZE:
+            logger.critical(
+                f"CORRUPTED WATCH DATA: Watch {uuid} file is unexpectedly large: "
+                f"{file_size / 1024 / 1024:.2f}MB (max: {MAX_WATCH_SIZE / 1024 / 1024}MB). "
+                f"File: {watch_json}. This indicates a bug or data corruption. "
+                f"Watch will be skipped."
+            )
+            return None
+
+        if HAS_ORJSON:
+            with open(watch_json, 'rb') as f:
+                watch_data = orjson.loads(f.read())
+        else:
+            with open(watch_json, 'r', encoding='utf-8') as f:
+                watch_data = json.load(f)
+
+        # Rehydrate and return watch object
+        watch_obj = rehydrate_entity_func(uuid, watch_data)
+        return watch_obj
+
+    except json.JSONDecodeError as e:
+        logger.critical(
+            f"CORRUPTED WATCH DATA: Failed to parse JSON for watch {uuid}. "
+            f"File: {watch_json}. Error: {e}. "
+            f"Watch will be skipped and may need manual recovery from backup."
+        )
+        return None
+    except ValueError as e:
+        # orjson raises ValueError for invalid JSON
+        if "invalid json" in str(e).lower() or HAS_ORJSON:
+            logger.critical(
+                f"CORRUPTED WATCH DATA: Failed to parse JSON for watch {uuid}. "
+                f"File: {watch_json}. Error: {e}. "
+                f"Watch will be skipped and may need manual recovery from backup."
+            )
+            return None
+        # Re-raise if it's not a JSON parsing error
+        raise
+    except FileNotFoundError:
+        logger.error(f"Watch file not found: {watch_json} for watch {uuid}")
+        return None
+    except Exception as e:
+        logger.error(f"Failed to load watch {uuid} from {watch_json}: {e}")
+        return None
+
+
+def load_all_watches(datastore_path, rehydrate_entity_func):
+    """
+    Load all watches from individual watch.json files.
+
+    SYNCHRONOUS loading: Blocks until all watches are loaded.
+    This ensures data consistency - web server won't accept requests
+    until all watches are available. Progress logged every 100 watches.
+
+    Args:
+        datastore_path: Path to the datastore directory
+        rehydrate_entity_func: Function to convert dict to Watch object
+
+    Returns:
+        Dictionary of uuid -> Watch object
+    """
+    start_time = time.time()
+    logger.info("Loading watches from individual watch.json files...")
+
+    watching = {}
+
+    if not os.path.exists(datastore_path):
+        return watching
+
+    # Find all watch.json files using glob (faster than manual directory traversal)
+    glob_start = time.time()
+    watch_files = glob.glob(os.path.join(datastore_path, "*", "watch.json"))
+    glob_time = time.time() - glob_start
+
+    total = len(watch_files)
+    logger.debug(f"Found {total} watch.json files in {glob_time:.3f}s")
+
+    loaded = 0
+    failed = 0
+
+    for watch_json in watch_files:
+        # Extract UUID from path: /datastore/{uuid}/watch.json
+        uuid_dir = os.path.basename(os.path.dirname(watch_json))
+        watch = load_watch_from_file(watch_json, uuid_dir, rehydrate_entity_func)
+        if watch:
+            watching[uuid_dir] = watch
+            loaded += 1
+
+            if loaded % 100 == 0:
+                logger.info(f"Loaded {loaded}/{total} watches...")
+        else:
+            # load_watch_from_file already logged the specific error
+            failed += 1
+
+    elapsed = time.time() - start_time
+
+    if failed > 0:
+        logger.critical(
+            f"LOAD COMPLETE: {loaded} watches loaded successfully, "
+            f"{failed} watches FAILED to load (corrupted or invalid) "
+            f"in {elapsed:.2f}s ({loaded/elapsed:.0f} watches/sec)"
+        )
+    else:
+        logger.info(f"Loaded {loaded} watches from disk in {elapsed:.2f}s ({loaded/elapsed:.0f} watches/sec)")
+
+    return watching
+
+
+def load_tag_from_file(tag_json, uuid, rehydrate_entity_func):
+    """
+    Load a tag from its JSON file.
+
+    Args:
+        tag_json: Path to the tag.json file
+        uuid: Tag UUID
+        rehydrate_entity_func: Function to convert dict to Tag object
+
+    Returns:
+        Tag object or None if failed
+    """
+    try:
+        # Check file size before reading
+        file_size = os.path.getsize(tag_json)
+        MAX_TAG_SIZE = 1 * 1024 * 1024  # 1MB
+        if file_size > MAX_TAG_SIZE:
+            logger.critical(
+                f"CORRUPTED TAG DATA: Tag {uuid} file is unexpectedly large: "
+                f"{file_size / 1024 / 1024:.2f}MB (max: {MAX_TAG_SIZE / 1024 / 1024}MB). "
+                f"File: {tag_json}. This indicates a bug or data corruption. "
+                f"Tag will be skipped."
+            )
+            return None
+
+        if HAS_ORJSON:
+            with open(tag_json, 'rb') as f:
+                tag_data = orjson.loads(f.read())
+        else:
+            with open(tag_json, 'r', encoding='utf-8') as f:
+                tag_data = json.load(f)
+
+        tag_data['processor'] = 'restock_diff'
+        # Rehydrate tag (convert dict to Tag object)
+        # processor_override is set inside the rehydration function
+        tag_obj = rehydrate_entity_func(uuid, tag_data)
+        return tag_obj
+
+    except json.JSONDecodeError as e:
+        logger.critical(
+            f"CORRUPTED TAG DATA: Failed to parse JSON for tag {uuid}. "
+            f"File: {tag_json}. Error: {e}. "
+            f"Tag will be skipped and may need manual recovery from backup."
+        )
+        return None
+    except ValueError as e:
+        # orjson raises ValueError for invalid JSON
+        if "invalid json" in str(e).lower() or HAS_ORJSON:
+            logger.critical(
+                f"CORRUPTED TAG DATA: Failed to parse JSON for tag {uuid}. "
+                f"File: {tag_json}. Error: {e}. "
+                f"Tag will be skipped and may need manual recovery from backup."
+            )
+            return None
+        # Re-raise if it's not a JSON parsing error
+        raise
+    except FileNotFoundError:
+        logger.debug(f"Tag file not found: {tag_json} for tag {uuid}")
+        return None
+    except Exception as e:
+        logger.error(f"Failed to load tag {uuid} from {tag_json}: {e}")
+        return None
+
+
+def load_all_tags(datastore_path, rehydrate_entity_func):
+    """
+    Load all tags from individual tag.json files.
+
+    Tags are stored separately from settings in {uuid}/tag.json files.
+
+    Args:
+        datastore_path: Path to the datastore directory
+        rehydrate_entity_func: Function to convert dict to Tag object
+
+    Returns:
+        Dictionary of uuid -> Tag object
+    """
+    logger.info("Loading tags from individual tag.json files...")
+
+    tags = {}
+
+    if not os.path.exists(datastore_path):
+        return tags
+
+    # Find all tag.json files using glob
+    tag_files = glob.glob(os.path.join(datastore_path, "*", "tag.json"))
+
+    total = len(tag_files)
+    if total == 0:
+        logger.debug("No tag.json files found")
+        return tags
+
+    logger.debug(f"Found {total} tag.json files")
+
+    loaded = 0
+    failed = 0
+
+    for tag_json in tag_files:
+        # Extract UUID from path: /datastore/{uuid}/tag.json
+        uuid_dir = os.path.basename(os.path.dirname(tag_json))
+        tag = load_tag_from_file(tag_json, uuid_dir, rehydrate_entity_func)
+        if tag:
+            tags[uuid_dir] = tag
+            loaded += 1
+        else:
+            # load_tag_from_file already logged the specific error
+            failed += 1
+
+    if failed > 0:
+        logger.warning(f"Loaded {loaded} tags, {failed} tags FAILED to load")
+    else:
+        logger.info(f"Loaded {loaded} tags from disk")
+
+    return tags
+
+
+# ============================================================================
+# FileSavingDataStore Class
+# ============================================================================
+
+class FileSavingDataStore(DataStore):
+    """
+    Abstract datastore that provides file persistence with immediate commits.
+
+    Features:
+    - Individual watch.json files (one per watch)
+    - Immediate persistence via watch.commit() and datastore.commit()
+    - Atomic file writes for crash safety
+
+    Subclasses must implement:
+    - rehydrate_entity(): Convert dict to Watch object
+    - Access to internal __data structure for watch management
+    """
+
+    def __init__(self):
+        super().__init__()
+
+    def _save_settings(self):
+        """
+        Save settings to storage (polymorphic).
+
+        Subclasses must implement for their backend.
+        - File: changedetection.json
+        - Redis: SET settings
+        - SQL: UPDATE settings table
+        """
+        raise NotImplementedError("Subclass must implement _save_settings")
+
+
+    def _load_watches(self):
+        """
+        Load all watches from storage (polymorphic).
+
+        Subclasses must implement for their backend.
+        - File: Read individual watch.json files
+        - Redis: SCAN watch:* keys
+        - SQL: SELECT * FROM watches
+        """
+        raise NotImplementedError("Subclass must implement _load_watches")
+
+    def _delete_watch(self, uuid):
+        """
+        Delete a watch from storage (polymorphic).
+
+        Subclasses must implement for their backend.
+        - File: Delete {uuid}/ directory recursively
+        - Redis: DEL watch:{uuid}
+        - SQL: DELETE FROM watches WHERE uuid=?
+
+        Args:
+            uuid: Watch UUID to delete
+        """
+        raise NotImplementedError("Subclass must implement _delete_watch")
+
+

changedetectionio/store/updates.py

@@ -0,0 +1,725 @@
+"""
+Schema update migrations for the datastore.
+
+This module contains all schema version upgrade methods (update_1 through update_N).
+These are mixed into ChangeDetectionStore to keep the main store file focused.
+
+IMPORTANT: Each update could be run even when they have a new install and the schema is correct.
+Therefore - each `update_n` should be very careful about checking if it needs to actually run.
+"""
+
+import os
+import re
+import shutil
+import tarfile
+import time
+from loguru import logger
+from copy import deepcopy
+
+
+# Try to import orjson for faster JSON serialization
+try:
+    import orjson
+    HAS_ORJSON = True
+except ImportError:
+    HAS_ORJSON = False
+
+from ..html_tools import TRANSLATE_WHITESPACE_TABLE
+from ..processors.restock_diff import Restock
+from ..blueprint.rss import RSS_CONTENT_FORMAT_DEFAULT
+from ..model import USE_SYSTEM_DEFAULT_NOTIFICATION_FORMAT_FOR_WATCH
+
+def create_backup_tarball(datastore_path, update_number):
+    """
+    Create a tarball backup of the entire datastore structure before running an update.
+
+    Includes:
+    - All {uuid}/watch.json files
+    - All {uuid}/tag.json files
+    - changedetection.json (settings, if it exists)
+    - url-watches.json (legacy format, if it exists)
+    - Directory structure preserved
+
+    Args:
+        datastore_path: Path to datastore directory
+        update_number: Update number being applied
+
+    Returns:
+        str: Path to created tarball, or None if backup failed
+
+    Restoration:
+    To restore from a backup:
+        cd /path/to/datastore
+        tar -xzf before-update-N-timestamp.tar.gz
+    This will restore all watch.json and tag.json files and settings to their pre-update state.
+    """
+    timestamp = int(time.time())
+    backup_filename = f"before-update-{update_number}-{timestamp}.tar.gz"
+    backup_path = os.path.join(datastore_path, backup_filename)
+
+    try:
+        logger.info(f"Creating backup tarball: {backup_filename}")
+
+        with tarfile.open(backup_path, "w:gz") as tar:
+            # Backup changedetection.json if it exists (new format)
+            changedetection_json = os.path.join(datastore_path, "changedetection.json")
+            if os.path.isfile(changedetection_json):
+                tar.add(changedetection_json, arcname="changedetection.json")
+                logger.debug("Added changedetection.json to backup")
+
+            # Backup url-watches.json if it exists (legacy format)
+            url_watches_json = os.path.join(datastore_path, "url-watches.json")
+            if os.path.isfile(url_watches_json):
+                tar.add(url_watches_json, arcname="url-watches.json")
+                logger.debug("Added url-watches.json to backup")
+
+            # Backup all watch/tag directories with their JSON files
+            # This preserves the UUID directory structure
+            watch_count = 0
+            tag_count = 0
+            for entry in os.listdir(datastore_path):
+                entry_path = os.path.join(datastore_path, entry)
+
+                # Skip if not a directory
+                if not os.path.isdir(entry_path):
+                    continue
+
+                # Skip hidden directories and backup directories
+                if entry.startswith('.') or entry.startswith('before-update-'):
+                    continue
+
+                # Backup watch.json if exists
+                watch_json = os.path.join(entry_path, "watch.json")
+                if os.path.isfile(watch_json):
+                    tar.add(watch_json, arcname=f"{entry}/watch.json")
+                    watch_count += 1
+
+                    if watch_count % 100 == 0:
+                        logger.debug(f"Backed up {watch_count} watch.json files...")
+
+                # Backup tag.json if exists
+                tag_json = os.path.join(entry_path, "tag.json")
+                if os.path.isfile(tag_json):
+                    tar.add(tag_json, arcname=f"{entry}/tag.json")
+                    tag_count += 1
+
+            logger.success(f"Backup created: {backup_filename} ({watch_count} watches from disk, {tag_count} tags from disk)")
+            return backup_path
+
+    except Exception as e:
+        logger.error(f"Failed to create backup tarball: {e}")
+        # Try to clean up partial backup
+        if os.path.exists(backup_path):
+            try:
+                os.unlink(backup_path)
+            except:
+                pass
+        return None
+
+
+class DatastoreUpdatesMixin:
+    """
+    Mixin class containing all schema update methods.
+
+    This class is inherited by ChangeDetectionStore to provide schema migration functionality.
+    Each update_N method upgrades the schema from version N-1 to version N.
+    """
+
+    def get_updates_available(self):
+        """
+        Discover all available update methods.
+
+        Returns:
+            list: Sorted list of update version numbers (e.g., [1, 2, 3, ..., 26])
+        """
+        import inspect
+        updates_available = []
+        for i, o in inspect.getmembers(self, predicate=inspect.ismethod):
+            m = re.search(r'update_(\d+)$', i)
+            if m:
+                updates_available.append(int(m.group(1)))
+        updates_available.sort()
+
+        return updates_available
+
+    def run_updates(self, current_schema_version=None):
+        import sys
+        """
+        Run all pending schema updates sequentially.
+
+        Args:
+            current_schema_version: Optional current schema version. If provided, only run updates
+                                   greater than this version. If None, uses the schema version from
+                                   the datastore. If no schema version exists in datastore and it appears
+                                   to be a fresh install, sets to latest update number (no updates needed).
+
+        IMPORTANT: Each update could be run even when they have a new install and the schema is correct.
+        Therefore - each `update_n` should be very careful about checking if it needs to actually run.
+
+        Process:
+        1. Get list of available updates
+        2. For each update > current schema version:
+           - Create backup of datastore
+           - Run update method
+           - Update schema version and commit settings
+           - Commit all watches and tags
+        3. If any update fails, stop processing
+        4. All changes saved via individual .commit() calls
+        """
+        updates_available = self.get_updates_available()
+        if self.data.get('watching'):
+            test_watch = self.data['watching'].get(next(iter(self.data.get('watching', {}))))
+            from ..model.Watch import model
+
+            if not isinstance(test_watch, model):
+                import sys
+                logger.critical("Cannot run updates! Watch structure must be re-hydrated back to a Watch model object!")
+                sys.exit(1)
+
+        if self.data['settings']['application'].get('tags',{}):
+            test_tag = self.data['settings']['application'].get('tags',{}).get(next(iter(self.data['settings']['application'].get('tags',{}))))
+            from ..model.Tag import model as tag_model
+
+            if not isinstance(test_tag, tag_model):
+                import sys
+                logger.critical("Cannot run updates! Watch tag/group structure must be re-hydrated back to a Tag model object!")
+                sys.exit(1)
+
+        # Determine current schema version
+        if current_schema_version is None:
+            # Check if schema_version exists in datastore
+            current_schema_version = self.data['settings']['application'].get('schema_version')
+
+            if current_schema_version is None:
+                # No schema version found - could be a fresh install or very old datastore
+                # If this is a fresh/new config with no watches, assume it's up-to-date
+                # and set to latest update number (no updates needed)
+                if len(self.data['watching']) == 0:
+                    # Get the highest update number from available update methods
+                    latest_update = updates_available[-1] if updates_available else 0
+                    logger.info(f"No schema version found and no watches exist - assuming fresh install, setting schema_version to {latest_update}")
+                    self.data['settings']['application']['schema_version'] = latest_update
+                    self.commit()
+                    return  # No updates needed for fresh install
+                else:
+                    # Has watches but no schema version - likely old datastore, run all updates
+                    logger.warning("No schema version found but watches exist - running all updates from version 0")
+                    current_schema_version = 0
+
+        logger.info(f"Current schema version: {current_schema_version}")
+
+        updates_ran = []
+
+        for update_n in updates_available:
+            if update_n > current_schema_version:
+                logger.critical(f"Applying update_{update_n}")
+
+                # Create tarball backup of entire datastore structure
+                # This includes all watch.json files, settings, and preserves directory structure
+                backup_path = create_backup_tarball(self.datastore_path, update_n)
+                if backup_path:
+                    logger.info(f"Backup created at: {backup_path}")
+                else:
+                    logger.warning("Backup creation failed, but continuing with update")
+
+                try:
+                    update_method = getattr(self, f"update_{update_n}")()
+                except Exception as e:
+                    logger.critical(f"Error while trying update_{update_n}")
+                    logger.exception(e)
+                    sys.exit(1)
+                else:
+                    # Bump the version
+                    self.data['settings']['application']['schema_version'] = update_n
+                    self.commit()
+
+                    logger.success(f"Update {update_n} completed")
+
+                    # Track which updates ran
+                    updates_ran.append(update_n)
+
+    # ============================================================================
+    # Individual Update Methods
+    # ============================================================================
+
+    def update_1(self):
+        """Convert minutes to seconds on settings and each watch."""
+        if self.data['settings']['requests'].get('minutes_between_check'):
+            self.data['settings']['requests']['time_between_check']['minutes'] = self.data['settings']['requests']['minutes_between_check']
+            # Remove the default 'hours' that is set from the model
+            self.data['settings']['requests']['time_between_check']['hours'] = None
+
+        for uuid, watch in self.data['watching'].items():
+            if 'minutes_between_check' in watch:
+                # Only upgrade individual watch time if it was set
+                if watch.get('minutes_between_check', False):
+                    self.data['watching'][uuid]['time_between_check']['minutes'] = watch['minutes_between_check']
+
+    def update_2(self):
+        """
+        Move the history list to a flat text file index.
+        Better than SQLite because this list is only appended to, and works across NAS / NFS type setups.
+        """
+        # @todo test running this on a newly updated one (when this already ran)
+        for uuid, watch in self.data['watching'].items():
+            history = []
+
+            if watch.get('history', False):
+                for d, p in watch['history'].items():
+                    d = int(d)  # Used to be keyed as str, we'll fix this now too
+                    history.append("{},{}\n".format(d, p))
+
+                if len(history):
+                    target_path = os.path.join(self.datastore_path, uuid)
+                    if os.path.exists(target_path):
+                        with open(os.path.join(target_path, "history.txt"), "w") as f:
+                            f.writelines(history)
+                    else:
+                        logger.warning(f"Datastore history directory {target_path} does not exist, skipping history import.")
+
+                # No longer needed, dynamically pulled from the disk when needed.
+                # But we should set it back to a empty dict so we don't break if this schema runs on an earlier version.
+                # In the distant future we can remove this entirely
+                self.data['watching'][uuid]['history'] = {}
+
+    def update_3(self):
+        """We incorrectly stored last_changed when there was not a change, and then confused the output list table."""
+        # see https://github.com/dgtlmoon/changedetection.io/pull/835
+        return
+
+    def update_4(self):
+        """`last_changed` not needed, we pull that information from the history.txt index."""
+        for uuid, watch in self.data['watching'].items():
+            try:
+                # Remove it from the struct
+                del(watch['last_changed'])
+            except:
+                continue
+        return
+
+    def update_5(self):
+        """
+        If the watch notification body, title look the same as the global one, unset it, so the watch defaults back to using the main settings.
+        In other words - the watch notification_title and notification_body are not needed if they are the same as the default one.
+        """
+        current_system_body = self.data['settings']['application']['notification_body'].translate(TRANSLATE_WHITESPACE_TABLE)
+        current_system_title = self.data['settings']['application']['notification_body'].translate(TRANSLATE_WHITESPACE_TABLE)
+        for uuid, watch in self.data['watching'].items():
+            try:
+                watch_body = watch.get('notification_body', '')
+                if watch_body and watch_body.translate(TRANSLATE_WHITESPACE_TABLE) == current_system_body:
+                    # Looks the same as the default one, so unset it
+                    watch['notification_body'] = None
+
+                watch_title = watch.get('notification_title', '')
+                if watch_title and watch_title.translate(TRANSLATE_WHITESPACE_TABLE) == current_system_title:
+                    # Looks the same as the default one, so unset it
+                    watch['notification_title'] = None
+            except Exception as e:
+                continue
+        return
+
+    def update_7(self):
+        """
+        We incorrectly used common header overrides that should only apply to Requests.
+        These are now handled in content_fetcher::html_requests and shouldnt be passed to Playwright/Selenium.
+        """
+        # These were hard-coded in early versions
+        for v in ['User-Agent', 'Accept', 'Accept-Encoding', 'Accept-Language']:
+            if self.data['settings']['headers'].get(v):
+                del self.data['settings']['headers'][v]
+
+    def update_8(self):
+        """Convert filters to a list of filters css_filter -> include_filters."""
+        for uuid, watch in self.data['watching'].items():
+            try:
+                existing_filter = watch.get('css_filter', '')
+                if existing_filter:
+                    watch['include_filters'] = [existing_filter]
+            except:
+                continue
+        return
+
+    def update_9(self):
+        """Convert old static notification tokens to jinja2 tokens."""
+        # Each watch
+        # only { } not {{ or }}
+        r = r'(?<!{){(?!{)(\w+)(?<!})}(?!})'
+        for uuid, watch in self.data['watching'].items():
+            try:
+                n_body = watch.get('notification_body', '')
+                if n_body:
+                    watch['notification_body'] = re.sub(r, r'{{\1}}', n_body)
+
+                n_title = watch.get('notification_title')
+                if n_title:
+                    watch['notification_title'] = re.sub(r, r'{{\1}}', n_title)
+
+                n_urls = watch.get('notification_urls')
+                if n_urls:
+                    for i, url in enumerate(n_urls):
+                        watch['notification_urls'][i] = re.sub(r, r'{{\1}}', url)
+
+            except:
+                continue
+
+        # System wide
+        n_body = self.data['settings']['application'].get('notification_body')
+        if n_body:
+            self.data['settings']['application']['notification_body'] = re.sub(r, r'{{\1}}', n_body)
+
+        n_title = self.data['settings']['application'].get('notification_title')
+        if n_body:
+            self.data['settings']['application']['notification_title'] = re.sub(r, r'{{\1}}', n_title)
+
+        n_urls = self.data['settings']['application'].get('notification_urls')
+        if n_urls:
+            for i, url in enumerate(n_urls):
+                self.data['settings']['application']['notification_urls'][i] = re.sub(r, r'{{\1}}', url)
+
+        return
+
+    def update_10(self):
+        """Some setups may have missed the correct default, so it shows the wrong config in the UI, although it will default to system-wide."""
+        for uuid, watch in self.data['watching'].items():
+            try:
+                if not watch.get('fetch_backend', ''):
+                    watch['fetch_backend'] = 'system'
+            except:
+                continue
+        return
+
+    def update_12(self):
+        """Create tag objects and their references from existing tag text."""
+        i = 0
+        for uuid, watch in self.data['watching'].items():
+            # Split out and convert old tag string
+            tag = watch.get('tag')
+            if tag:
+                tag_uuids = []
+                for t in tag.split(','):
+                    tag_uuids.append(self.add_tag(title=t))
+
+                self.data['watching'][uuid]['tags'] = tag_uuids
+
+    def update_13(self):
+        """#1775 - Update 11 did not update the records correctly when adding 'date_created' values for sorting."""
+        i = 0
+        for uuid, watch in self.data['watching'].items():
+            if not watch.get('date_created'):
+                self.data['watching'][uuid]['date_created'] = i
+            i += 1
+        return
+
+    def update_14(self):
+        """#1774 - protect xpath1 against migration."""
+        for awatch in self.data["watching"]:
+            if self.data["watching"][awatch]['include_filters']:
+                for num, selector in enumerate(self.data["watching"][awatch]['include_filters']):
+                    if selector.startswith('/'):
+                        self.data["watching"][awatch]['include_filters'][num] = 'xpath1:' + selector
+                    if selector.startswith('xpath:'):
+                        self.data["watching"][awatch]['include_filters'][num] = selector.replace('xpath:', 'xpath1:', 1)
+
+    def update_15(self):
+        """Use more obvious default time setting."""
+        for uuid in self.data["watching"]:
+            if self.data["watching"][uuid]['time_between_check'] == self.data['settings']['requests']['time_between_check']:
+                # What the old logic was, which was pretty confusing
+                self.data["watching"][uuid]['time_between_check_use_default'] = True
+            elif all(value is None or value == 0 for value in self.data["watching"][uuid]['time_between_check'].values()):
+                self.data["watching"][uuid]['time_between_check_use_default'] = True
+            else:
+                # Something custom here
+                self.data["watching"][uuid]['time_between_check_use_default'] = False
+
+    def update_16(self):
+        """Correctly set datatype for older installs where 'tag' was string and update_12 did not catch it."""
+        for uuid, watch in self.data['watching'].items():
+            if isinstance(watch.get('tags'), str):
+                self.data['watching'][uuid]['tags'] = []
+
+    def update_17(self):
+        """Migrate old 'in_stock' values to the new Restock."""
+        for uuid, watch in self.data['watching'].items():
+            if 'in_stock' in watch:
+                watch['restock'] = Restock({'in_stock': watch.get('in_stock')})
+                del watch['in_stock']
+
+    def update_18(self):
+        """Migrate old restock settings."""
+        for uuid, watch in self.data['watching'].items():
+            if not watch.get('restock_settings'):
+                # So we enable price following by default
+                self.data['watching'][uuid]['restock_settings'] = {'follow_price_changes': True}
+
+            # Migrate and cleanoff old value
+            self.data['watching'][uuid]['restock_settings']['in_stock_processing'] = 'in_stock_only' if watch.get(
+                'in_stock_only') else 'all_changes'
+
+            if self.data['watching'][uuid].get('in_stock_only'):
+                del (self.data['watching'][uuid]['in_stock_only'])
+
+    def update_19(self):
+        """Compress old elements.json to elements.deflate, saving disk, this compression is pretty fast."""
+        import zlib
+
+        for uuid, watch in self.data['watching'].items():
+            json_path = os.path.join(self.datastore_path, uuid, "elements.json")
+            deflate_path = os.path.join(self.datastore_path, uuid, "elements.deflate")
+
+            if os.path.exists(json_path):
+                with open(json_path, "rb") as f_j:
+                    with open(deflate_path, "wb") as f_d:
+                        logger.debug(f"Compressing {str(json_path)} to {str(deflate_path)}..")
+                        f_d.write(zlib.compress(f_j.read()))
+                        os.unlink(json_path)
+
+    def update_20(self):
+        """Migrate extract_title_as_title to use_page_title_in_list."""
+        for uuid, watch in self.data['watching'].items():
+            if self.data['watching'][uuid].get('extract_title_as_title'):
+                self.data['watching'][uuid]['use_page_title_in_list'] = self.data['watching'][uuid].get('extract_title_as_title')
+                del self.data['watching'][uuid]['extract_title_as_title']
+
+        if self.data['settings']['application'].get('extract_title_as_title'):
+            # Ensure 'ui' key exists (defensive for edge cases where base_config merge didn't happen)
+            if 'ui' not in self.data['settings']['application']:
+                self.data['settings']['application']['ui'] = {
+                    'use_page_title_in_list': True,
+                    'open_diff_in_new_tab': True,
+                    'socket_io_enabled': True,
+                    'favicons_enabled': True
+                }
+            self.data['settings']['application']['ui']['use_page_title_in_list'] = self.data['settings']['application'].get('extract_title_as_title')
+
+    def update_21(self):
+        """Migrate timezone to scheduler_timezone_default."""
+        if self.data['settings']['application'].get('timezone'):
+            self.data['settings']['application']['scheduler_timezone_default'] = self.data['settings']['application'].get('timezone')
+            del self.data['settings']['application']['timezone']
+
+    def update_23(self):
+        """Some notification formats got the wrong name type."""
+
+        def re_run(formats):
+            sys_n_format = self.data['settings']['application'].get('notification_format')
+            key_exists_as_value = next((k for k, v in formats.items() if v == sys_n_format), None)
+            if key_exists_as_value:  # key of "Plain text"
+                logger.success(f"['settings']['application']['notification_format'] '{sys_n_format}' -> '{key_exists_as_value}'")
+                self.data['settings']['application']['notification_format'] = key_exists_as_value
+
+            for uuid, watch in self.data['watching'].items():
+                n_format = self.data['watching'][uuid].get('notification_format')
+                key_exists_as_value = next((k for k, v in formats.items() if v == n_format), None)
+                if key_exists_as_value and key_exists_as_value != USE_SYSTEM_DEFAULT_NOTIFICATION_FORMAT_FOR_WATCH:  # key of "Plain text"
+                    logger.success(f"['watching'][{uuid}]['notification_format'] '{n_format}' -> '{key_exists_as_value}'")
+                    self.data['watching'][uuid]['notification_format'] = key_exists_as_value  # should be 'text' or whatever
+
+            for uuid, tag in self.data['settings']['application']['tags'].items():
+                n_format = self.data['settings']['application']['tags'][uuid].get('notification_format')
+                key_exists_as_value = next((k for k, v in formats.items() if v == n_format), None)
+                if key_exists_as_value and key_exists_as_value != USE_SYSTEM_DEFAULT_NOTIFICATION_FORMAT_FOR_WATCH:  # key of "Plain text"
+                    logger.success(
+                        f"['settings']['application']['tags'][{uuid}]['notification_format'] '{n_format}' -> '{key_exists_as_value}'")
+                    self.data['settings']['application']['tags'][uuid][
+                        'notification_format'] = key_exists_as_value  # should be 'text' or whatever
+
+        from ..notification import valid_notification_formats
+        formats = deepcopy(valid_notification_formats)
+        re_run(formats)
+        # And in previous versions, it was "text" instead of Plain text, Markdown instead of "Markdown to HTML"
+        formats['text'] = 'Text'
+        formats['markdown'] = 'Markdown'
+        re_run(formats)
+
+    def update_24(self):
+        """RSS types should be inline with the same names as notification types."""
+        rss_format = self.data['settings']['application'].get('rss_content_format')
+        if not rss_format or 'text' in rss_format:
+            # might have been 'plaintext, 'plain text' or something
+            self.data['settings']['application']['rss_content_format'] = RSS_CONTENT_FORMAT_DEFAULT
+        elif 'html' in rss_format:
+            self.data['settings']['application']['rss_content_format'] = 'htmlcolor'
+        else:
+            # safe fallback to text
+            self.data['settings']['application']['rss_content_format'] = RSS_CONTENT_FORMAT_DEFAULT
+
+    def update_25(self):
+        """Different processors now hold their own history.txt."""
+        for uuid, watch in self.data['watching'].items():
+            processor = self.data['watching'][uuid].get('processor')
+            if processor != 'text_json_diff':
+                old_history_txt = os.path.join(self.datastore_path, "history.txt")
+                target_history_name = f"history-{processor}.txt"
+                if os.path.isfile(old_history_txt) and not os.path.isfile(target_history_name):
+                    new_history_txt = os.path.join(self.datastore_path, target_history_name)
+                    logger.debug(f"Renaming history index {old_history_txt} to {new_history_txt}...")
+                    shutil.move(old_history_txt, new_history_txt)
+
+    def migrate_legacy_db_format(self):
+        """
+        Migration: Individual watch persistence (COPY-based, safe rollback).
+
+        Loads legacy url-watches.json format and migrates to:
+        - {uuid}/watch.json (per watch)
+        - changedetection.json (settings only)
+
+        IMPORTANT:
+        - A tarball backup (before-update-26-timestamp.tar.gz) is created before migration
+        - url-watches.json is LEFT INTACT for rollback safety
+        - Users can roll back by simply downgrading to the previous version
+        - Or restore from tarball: tar -xzf before-update-26-*.tar.gz
+
+        This is a dedicated migration release - users upgrade at their own pace.
+        """
+        logger.critical("=" * 80)
+        logger.critical("Running migration: Individual watch persistence (update_26)")
+        logger.critical("COPY-based migration: url-watches.json will remain intact for rollback")
+        logger.critical("=" * 80)
+
+        # Populate settings from legacy data
+        logger.info("Populating settings from legacy data...")
+        watch_count = len(self.data['watching'])
+        logger.success(f"Loaded {watch_count} watches from legacy format")
+
+        # Phase 1: Save all watches to individual files
+        logger.critical(f"Phase 1/4: Saving {watch_count} watches to individual watch.json files...")
+
+        saved_count = 0
+        for uuid, watch in self.data['watching'].items():
+            try:
+                watch.commit()
+                saved_count += 1
+
+                if saved_count % 100 == 0:
+                    logger.info(f"  Progress: {saved_count}/{watch_count} watches migrated...")
+
+            except Exception as e:
+                logger.error(f"Failed to save watch {uuid}: {e}")
+                raise Exception(
+                    f"Migration failed: Could not save watch {uuid}. "
+                    f"url-watches.json remains intact, safe to retry. Error: {e}"
+                )
+
+        logger.critical(f"Phase 1 complete: Saved {saved_count} watches")
+
+        # Phase 2: Verify all files exist
+        logger.critical("Phase 2/4: Verifying all watch.json files were created...")
+
+        missing = []
+        for uuid in self.data['watching'].keys():
+            watch_json = os.path.join(self.datastore_path, uuid, "watch.json")
+            if not os.path.isfile(watch_json):
+                missing.append(uuid)
+
+        if missing:
+            raise Exception(
+                f"Migration failed: {len(missing)} watch files missing: {missing[:5]}... "
+                f"url-watches.json remains intact, safe to retry."
+            )
+
+        logger.critical(f"Phase 2 complete: Verified {watch_count} watch files")
+
+        # Phase 3: Create new settings file
+        logger.critical("Phase 3/4: Creating changedetection.json...")
+
+        try:
+            self._save_settings()
+        except Exception as e:
+            logger.error(f"Failed to create changedetection.json: {e}")
+            raise Exception(
+                f"Migration failed: Could not create changedetection.json. "
+                f"url-watches.json remains intact, safe to retry. Error: {e}"
+            )
+
+        # Phase 4: Verify settings file exists
+        logger.critical("Phase 4/4: Verifying changedetection.json exists...")
+        changedetection_json_new_schema=os.path.join(self.datastore_path, "changedetection.json")
+        if not os.path.isfile(changedetection_json_new_schema):
+            import sys
+            logger.critical("Migration failed, changedetection.json not found after update ran!")
+            sys.exit(1)
+
+
+        logger.critical("Phase 4 complete: Verified changedetection.json exists")
+
+        # Success! Now reload from new format
+        logger.critical("Reloading datastore from new format...")
+        # write it to disk, it will be saved without ['watching'] in the JSON db because we find it from disk glob
+        self._save_settings()
+        logger.success("Datastore reloaded from new format successfully")
+        logger.critical("=" * 80)
+        logger.critical("MIGRATION COMPLETED SUCCESSFULLY!")
+        logger.critical("=" * 80)
+        logger.info("")
+        logger.info("New format:")
+        logger.info(f"  - {watch_count} individual watch.json files created")
+        logger.info(f"  - changedetection.json created (settings only)")
+        logger.info("")
+        logger.info("Rollback safety:")
+        logger.info("  - url-watches.json preserved for rollback")
+        logger.info("  - To rollback: downgrade to previous version and restart")
+        logger.info("  - No manual file operations needed")
+        logger.info("")
+        logger.info("Optional cleanup (after testing new version):")
+        logger.info(f"  - rm {os.path.join(self.datastore_path, 'url-watches.json')}")
+        logger.info("")
+
+    def update_26(self):
+        self.migrate_legacy_db_format()
+
+    def update_28(self):
+        """
+        Migrate tags to individual tag.json files.
+
+        Tags are currently saved only in changedetection.json (settings).
+        This migration ALSO saves them to individual {uuid}/tag.json files,
+        similar to how watches are stored (dual storage).
+
+        Benefits:
+        - Allows atomic tag updates without rewriting entire settings
+        - Enables independent tag versioning/backup
+        - Maintains backwards compatibility (tags stay in settings too)
+        """
+        # Force save as tag.json (not watch.json) even if object is corrupted
+
+        logger.critical("=" * 80)
+        logger.critical("Running migration: Individual tag persistence (update_28)")
+        logger.critical("Creating individual tag.json files")
+        logger.critical("=" * 80)
+
+        tags = self.data['settings']['application'].get('tags', {})
+        tag_count = len(tags)
+
+        if tag_count == 0:
+            logger.info("No tags found, skipping migration")
+            return
+
+        logger.info(f"Migrating {tag_count} tags to individual tag.json files...")
+
+        saved_count = 0
+        failed_count = 0
+
+        for uuid, tag_data in tags.items():
+            try:
+                tag_data.commit()
+                saved_count += 1
+                if saved_count % 10 == 0:
+                    logger.info(f"  Progress: {saved_count}/{tag_count} tags migrated...")
+
+            except Exception as e:
+                logger.error(f"Failed to save tag {uuid} ({tag_data.get('title', 'unknown')}): {e}")
+                failed_count += 1
+
+        if failed_count > 0:
+            logger.warning(f"Migration complete: {saved_count} tags saved, {failed_count} tags FAILED")
+        else:
+            logger.success(f"Migration complete: {saved_count} tags saved to individual tag.json files")
+
+        # Tags remain in settings for backwards compatibility AND easy access
+        # On next load, _load_tags() will read from tag.json files and merge with settings
+        logger.info("Tags saved to both settings AND individual tag.json files")
+        logger.info("Future tag edits will update both locations (dual storage)")
+        logger.critical("=" * 80)
+

changedetectionio/templates/base.html

@@ -27,7 +27,7 @@
     <link rel="apple-touch-icon" sizes="180x180" href="{{url_for('static_content', group='favicons', filename='apple-touch-icon.png')}}">
     <link rel="icon" type="image/png" sizes="32x32" href="{{url_for('static_content', group='favicons', filename='favicon-32x32.png')}}">
     <link rel="icon" type="image/png" sizes="16x16" href="{{url_for('static_content', group='favicons', filename='favicon-16x16.png')}}">
-    <link rel="manifest" href="{{url_for('static_content', group='favicons', filename='site.webmanifest')}}">
+    <link rel="manifest" href="{{url_for('static_content', group='favicons', filename='site.webmanifest')}}" crossorigin="use-credentials">
     <link rel="mask-icon" href="{{url_for('static_content', group='favicons', filename='safari-pinned-tab.svg')}}" color="#5bbad5">
     <link rel="shortcut icon" href="{{url_for('static_content', group='favicons', filename='favicon.ico')}}">
     <meta name="msapplication-TileColor" content="#da532c">
@@ -265,6 +265,9 @@
           </a>
           {% endfor %}
         </div>
+        <div>
+            <a href="{{ url_for('ui.delete_locale_language_session_var_if_it_exists', redirect=request.path) }}" >{{ _('Auto-detect from browser') }}</a>
+        </div>
         <div>
             {{ _('Language support is in beta, please help us improve by opening a PR on GitHub with any updates.') }}
         </div>

changedetectionio/test_cli_opts.sh

@@ -0,0 +1,265 @@
+#!/bin/bash
+# Test script for CLI options - Parallel execution
+# Tests -u, -uN, -r, -b flags
+
+set -u  # Exit on undefined variables
+
+# Color output
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+NC='\033[0m' # No Color
+
+# Test results directory (for parallel safety)
+TEST_RESULTS_DIR="/tmp/cli-test-results-$$"
+mkdir -p "$TEST_RESULTS_DIR"
+
+# Cleanup function
+cleanup() {
+    echo ""
+    echo "=== Cleaning up test directories ==="
+    rm -rf /tmp/cli-test-* 2>/dev/null || true
+    rm -rf "$TEST_RESULTS_DIR" 2>/dev/null || true
+    # Kill any hanging processes
+    pkill -f "changedetection.py.*cli-test" 2>/dev/null || true
+}
+trap cleanup EXIT
+
+# Helper to record test result
+record_result() {
+    local test_num=$1
+    local status=$2  # pass or fail
+    local message=$3
+
+    echo "$status|$message" > "$TEST_RESULTS_DIR/test_${test_num}.result"
+}
+
+# Run a test in background
+run_test() {
+    local test_num=$1
+    local test_name=$2
+    local test_func=$3
+
+    (
+        echo -e "${YELLOW}[Test $test_num]${NC} $test_name"
+        if $test_func "$test_num"; then
+            record_result "$test_num" "pass" "$test_name"
+            echo -e "${GREEN}✓ PASS${NC}: $test_name"
+        else
+            record_result "$test_num" "fail" "$test_name"
+            echo -e "${RED}✗ FAIL${NC}: $test_name"
+        fi
+    ) &
+}
+
+# =============================================================================
+# Test Functions (each runs independently)
+# =============================================================================
+
+test_help_flag() {
+    local test_id=$1
+    timeout 3 python3 changedetection.py --help 2>&1 | grep -q "Add URLs on startup"
+}
+
+test_version_flag() {
+    local test_id=$1
+    timeout 3 python3 changedetection.py --version 2>&1 | grep -qE "changedetection.io [0-9]+\.[0-9]+"
+}
+
+test_single_url() {
+    local test_id=$1
+    local dir="/tmp/cli-test-single-${test_id}-$$"
+    timeout 10 python3 changedetection.py -d "$dir" -C -u https://example.com -b &>/dev/null
+    # Count watch directories (UUID directories containing watch.json)
+    [ "$(find "$dir" -mindepth 2 -maxdepth 2 -name 'watch.json' | wc -l)" -eq 1 ]
+}
+
+test_multiple_urls() {
+    local test_id=$1
+    local dir="/tmp/cli-test-multi-${test_id}-$$"
+    timeout 12 python3 changedetection.py -d "$dir" -C \
+        -u https://example.com \
+        -u https://github.com \
+        -u https://httpbin.org \
+        -b &>/dev/null
+    # Count watch directories (UUID directories containing watch.json)
+    [ "$(find "$dir" -mindepth 2 -maxdepth 2 -name 'watch.json' | wc -l)" -eq 3 ]
+}
+
+test_url_with_options() {
+    local test_id=$1
+    local dir="/tmp/cli-test-opts-${test_id}-$$"
+    timeout 10 python3 changedetection.py -d "$dir" -C \
+        -u https://example.com \
+        -u0 '{"title":"Test Site","processor":"text_json_diff"}' \
+        -b &>/dev/null
+    # Check that at least one watch.json contains the title "Test Site"
+    python3 -c "
+import json, glob, sys
+watch_files = glob.glob('$dir/*/watch.json')
+for wf in watch_files:
+    with open(wf) as f:
+        data = json.load(f)
+        if data.get('title') == 'Test Site':
+            sys.exit(0)
+sys.exit(1)
+"
+}
+
+test_multiple_urls_with_options() {
+    local test_id=$1
+    local dir="/tmp/cli-test-multi-opts-${test_id}-$$"
+    timeout 12 python3 changedetection.py -d "$dir" -C \
+        -u https://example.com \
+        -u0 '{"title":"Site One"}' \
+        -u https://github.com \
+        -u1 '{"title":"Site Two"}' \
+        -b &>/dev/null
+    # Check that we have 2 watches and both titles are present
+    python3 -c "
+import json, glob, sys
+watch_files = glob.glob('$dir/*/watch.json')
+if len(watch_files) != 2:
+    sys.exit(1)
+titles = []
+for wf in watch_files:
+    with open(wf) as f:
+        data = json.load(f)
+        titles.append(data.get('title'))
+sys.exit(0 if 'Site One' in titles and 'Site Two' in titles else 1)
+"
+}
+
+test_batch_mode_exit() {
+    local test_id=$1
+    local dir="/tmp/cli-test-batch-${test_id}-$$"
+    local start=$(date +%s)
+    timeout 15 python3 changedetection.py -d "$dir" -C \
+        -u https://example.com \
+        -b &>/dev/null
+    local end=$(date +%s)
+    local elapsed=$((end - start))
+    [ $elapsed -lt 14 ]
+}
+
+test_recheck_all() {
+    local test_id=$1
+    local dir="/tmp/cli-test-recheck-all-${test_id}-$$"
+    # Create a watch using CLI, then recheck it
+    timeout 10 python3 changedetection.py -d "$dir" -C -u https://example.com -b &>/dev/null
+    # Now recheck all watches
+    timeout 10 python3 changedetection.py -d "$dir" -r all -b 2>&1 | grep -q "Queuing"
+}
+
+test_recheck_specific() {
+    local test_id=$1
+    local dir="/tmp/cli-test-recheck-uuid-${test_id}-$$"
+    # Create 2 watches using CLI
+    timeout 12 python3 changedetection.py -d "$dir" -C \
+        -u https://example.com \
+        -u https://github.com \
+        -b &>/dev/null
+    # Get the UUIDs that were created
+    local uuids=$(find "$dir" -mindepth 2 -maxdepth 2 -name 'watch.json' -exec dirname {} \; | xargs -n1 basename | tr '\n' ',' | sed 's/,$//')
+    # Now recheck specific UUIDs
+    timeout 10 python3 changedetection.py -d "$dir" -r "$uuids" -b 2>&1 | grep -q "Queuing"
+}
+
+test_combined_operations() {
+    local test_id=$1
+    local dir="/tmp/cli-test-combined-${test_id}-$$"
+    timeout 12 python3 changedetection.py -d "$dir" -C \
+        -u https://example.com \
+        -u https://github.com \
+        -r all \
+        -b &>/dev/null
+    # Count watch directories (UUID directories containing watch.json)
+    [ "$(find "$dir" -mindepth 2 -maxdepth 2 -name 'watch.json' | wc -l)" -eq 2 ]
+}
+
+test_invalid_json() {
+    local test_id=$1
+    local dir="/tmp/cli-test-invalid-${test_id}-$$"
+    timeout 5 python3 changedetection.py -d "$dir" -C \
+        -u https://example.com \
+        -u0 'invalid json here' \
+        2>&1 | grep -qi "invalid json\|json decode error"
+}
+
+test_create_directory() {
+    local test_id=$1
+    local dir="/tmp/cli-test-create-${test_id}-$$/nested/path"
+    timeout 10 python3 changedetection.py -d "$dir" -C \
+        -u https://example.com \
+        -b &>/dev/null
+    [ -d "$dir" ]
+}
+
+# =============================================================================
+# Main Test Execution
+# =============================================================================
+
+echo "=========================================="
+echo "  CLI Options Test Suite (Parallel)"
+echo "=========================================="
+echo ""
+
+# Launch all tests in parallel
+run_test 1 "Help flag (--help) shows usage without initialization" test_help_flag
+run_test 2 "Version flag (--version) displays version" test_version_flag
+run_test 3 "Add single URL with -u flag" test_single_url
+run_test 4 "Add multiple URLs with multiple -u flags" test_multiple_urls
+run_test 5 "Add URL with JSON options using -u0" test_url_with_options
+run_test 6 "Add multiple URLs with different options (-u0, -u1)" test_multiple_urls_with_options
+run_test 7 "Batch mode (-b) exits automatically after processing" test_batch_mode_exit
+run_test 8 "Recheck all watches with -r all" test_recheck_all
+run_test 9 "Recheck specific watches with -r UUID" test_recheck_specific
+run_test 10 "Combined: Add URLs and recheck all with -u and -r all" test_combined_operations
+run_test 11 "Invalid JSON in -u0 option should show error" test_invalid_json
+run_test 12 "Create datastore directory with -C flag" test_create_directory
+
+# Wait for all tests to complete
+echo ""
+echo "Waiting for all tests to complete..."
+wait
+
+# Collect results
+echo ""
+echo "=========================================="
+echo "  Test Summary"
+echo "=========================================="
+
+TESTS_RUN=0
+TESTS_PASSED=0
+TESTS_FAILED=0
+
+for result_file in "$TEST_RESULTS_DIR"/test_*.result; do
+    if [ -f "$result_file" ]; then
+        TESTS_RUN=$((TESTS_RUN + 1))
+        status=$(cut -d'|' -f1 < "$result_file")
+        if [ "$status" = "pass" ]; then
+            TESTS_PASSED=$((TESTS_PASSED + 1))
+        else
+            TESTS_FAILED=$((TESTS_FAILED + 1))
+        fi
+    fi
+done
+
+echo "Tests run:    $TESTS_RUN"
+echo -e "${GREEN}Tests passed: $TESTS_PASSED${NC}"
+if [ $TESTS_FAILED -gt 0 ]; then
+    echo -e "${RED}Tests failed: $TESTS_FAILED${NC}"
+else
+    echo -e "${GREEN}Tests failed: $TESTS_FAILED${NC}"
+fi
+echo "=========================================="
+echo ""
+
+# Exit with appropriate code
+if [ $TESTS_FAILED -gt 0 ]; then
+    echo -e "${RED}Some tests failed!${NC}"
+    exit 1
+else
+    echo -e "${GREEN}All tests passed!${NC}"
+    exit 0
+fi

changedetectionio/tests/conftest.py

@@ -5,13 +5,16 @@ from threading import Thread
 
 import pytest
 import arrow
-from changedetectionio import changedetection_app
 from changedetectionio import store
 import os
 import sys
-from loguru import logger
 
-from changedetectionio.flask_app import init_app_secret
+# CRITICAL: Set short timeout for tests to prevent 45-second hangs
+# When test server is slow/unresponsive, workers fail fast instead of holding UUIDs for 45s
+# This prevents exponential priority growth from repeated deferrals (priority × 10 each defer)
+os.environ['DEFAULT_SETTINGS_REQUESTS_TIMEOUT'] = '5'
+
+from changedetectionio.flask_app import init_app_secret, changedetection_app
 from changedetectionio.tests.util import live_server_setup, new_live_server_setup
 
 # https://github.com/pallets/flask/blob/1.1.2/examples/tutorial/tests/test_auth.py
@@ -31,6 +34,93 @@ def reportlog(pytestconfig):
     logger.remove(handler_id)
 
 
+@pytest.fixture(autouse=True)
+def per_test_log_file(request):
+    """Create a separate log file for each test function with pytest output."""
+    import re
+
+    # Create logs directory if it doesn't exist
+    log_dir = os.path.join(os.path.dirname(__file__), "logs")
+    os.makedirs(log_dir, exist_ok=True)
+
+    # Generate log filename from test name and worker ID (for parallel runs)
+    test_name = request.node.name
+
+    # Sanitize test name - replace unsafe characters with underscores
+    # Keep only alphanumeric, dash, underscore, and period
+    safe_test_name = re.sub(r'[^\w\-.]', '_', test_name)
+
+    # Limit length to avoid filesystem issues (max 200 chars)
+    if len(safe_test_name) > 200:
+        # Keep first 150 chars + hash of full name + last 30 chars
+        import hashlib
+        name_hash = hashlib.md5(test_name.encode()).hexdigest()[:8]
+        safe_test_name = f"{safe_test_name[:150]}_{name_hash}_{safe_test_name[-30:]}"
+
+    worker_id = os.environ.get('PYTEST_XDIST_WORKER', 'master')
+    log_file = os.path.join(log_dir, f"{safe_test_name}_{worker_id}.log")
+
+    # Add file handler for this test with TRACE level
+    handler_id = logger.add(
+        log_file,
+        format="{time:YYYY-MM-DD HH:mm:ss.SSS} | {level: <8} | {process} | {name}:{function}:{line} - {message}",
+        level="TRACE",
+        mode="w",  # Overwrite if exists
+        enqueue=True  # Thread-safe
+    )
+
+    logger.info(f"=== Starting test: {test_name} (worker: {worker_id}) ===")
+    logger.info(f"Test location: {request.node.nodeid}")
+
+    yield
+
+    # Capture test outcome (PASSED/FAILED/SKIPPED/ERROR)
+    outcome = "UNKNOWN"
+    exc_info = None
+    stdout = None
+    stderr = None
+
+    if hasattr(request.node, 'rep_call'):
+        outcome = request.node.rep_call.outcome.upper()
+        if request.node.rep_call.failed:
+            exc_info = request.node.rep_call.longreprtext
+        # Capture stdout/stderr from call phase
+        if hasattr(request.node.rep_call, 'sections'):
+            for section_name, section_content in request.node.rep_call.sections:
+                if 'stdout' in section_name.lower():
+                    stdout = section_content
+                elif 'stderr' in section_name.lower():
+                    stderr = section_content
+    elif hasattr(request.node, 'rep_setup'):
+        if request.node.rep_setup.failed:
+            outcome = "SETUP_FAILED"
+            exc_info = request.node.rep_setup.longreprtext
+
+    logger.info(f"=== Test Result: {outcome} ===")
+
+    if exc_info:
+        logger.error(f"=== Test Failure Details ===\n{exc_info}")
+
+    if stdout:
+        logger.info(f"=== Captured stdout ===\n{stdout}")
+
+    if stderr:
+        logger.warning(f"=== Captured stderr ===\n{stderr}")
+
+    logger.info(f"=== Finished test: {test_name} ===")
+    logger.remove(handler_id)
+
+
+@pytest.hookimpl(tryfirst=True, hookwrapper=True)
+def pytest_runtest_makereport(item, call):
+    """Hook to capture test results and attach to the test node."""
+    outcome = yield
+    rep = outcome.get_result()
+
+    # Store report on the test node for access in fixtures
+    setattr(item, f"rep_{rep.when}", rep)
+
+
 @pytest.fixture
 def environment(mocker):
     """Mock arrow.now() to return a fixed datetime for testing jinja2 time extension."""
@@ -167,9 +257,56 @@ def prepare_test_function(live_server, datastore_path):
         except:
             break
 
-    # Prevent background thread from writing during cleanup/reload
-    datastore.needs_write = False
-    datastore.needs_write_urgent = False
+    # Add test helper methods to the app for worker management
+    def set_workers(count):
+        """Set the number of workers for testing - brutal shutdown, no delays"""
+        from changedetectionio import worker_pool
+        from changedetectionio.flask_app import update_q, notification_q
+
+        current_count = worker_pool.get_worker_count()
+
+       # Special case: Setting to 0 means shutdown all workers brutally
+        if count == 0:
+            logger.debug(f"Brutally shutting down all {current_count} workers")
+            worker_pool.shutdown_workers()
+            return {
+                'status': 'success',
+                'message': f'Shutdown all {current_count} workers',
+                'previous_count': current_count,
+                'current_count': 0
+            }
+
+        # Adjust worker count (no delays, no verification)
+        result = worker_pool.adjust_async_worker_count(
+            count,
+            update_q=update_q,
+            notification_q=notification_q,
+            app=live_server.app,
+            datastore=datastore
+        )
+
+        return result
+
+    def check_all_workers_alive(expected_count):
+        """Check that all expected workers are alive"""
+        from changedetectionio import worker_pool
+        from changedetectionio.flask_app import update_q, notification_q
+        result = worker_pool.check_worker_health(
+            expected_count,
+            update_q=update_q,
+            notification_q=notification_q,
+            app=live_server.app,
+            datastore=datastore
+        )
+        assert result['status'] == 'healthy', f"Workers not healthy: {result['message']}"
+        return result
+
+    # Attach helper methods to app for easy test access
+    live_server.app.set_workers = set_workers
+    live_server.app.check_all_workers_alive = check_all_workers_alive
+
+
+
 
     # CRITICAL: Clean up any files from previous tests
     # This ensures a completely clean directory
@@ -203,7 +340,6 @@ def prepare_test_function(live_server, datastore_path):
                 break
 
         datastore.data['watching'] = {}
-        datastore.needs_write = True
     except Exception as e:
         logger.warning(f"Error during datastore cleanup: {e}")
 
@@ -264,8 +400,8 @@ def app(request, datastore_path):
 
         # Shutdown workers gracefully before loguru cleanup
         try:
-            from changedetectionio import worker_handler
-            worker_handler.shutdown_workers()
+            from changedetectionio import worker_pool
+            worker_pool.shutdown_workers()
         except Exception:
             pass
 
@@ -313,4 +449,3 @@ def app(request, datastore_path):
     yield app
 
 
-

changedetectionio/tests/plugins/test_processor.py

@@ -0,0 +1,41 @@
+import time
+
+from flask import url_for
+
+from changedetectionio.tests.util import wait_for_all_checks
+
+
+def test_check_plugin_processor(client, live_server, measure_memory_usage, datastore_path):
+    # requires os-int intelligence plugin installed (first basic one we test with)
+
+    res = client.get(url_for("watchlist.index"))
+    assert b'OSINT Reconnaissance' in res.data, "Must have the OSINT plugin installed at test time"
+    assert b'<input checked id="processor-0" name="processor" type="radio" value="text_json_diff">' in res.data, "But the first text_json_diff processor should always be selected by default in quick watch form"
+
+    res = client.post(
+        url_for("ui.ui_views.form_quick_watch_add"),
+        data={"url": 'http://127.0.0.1', "tags": '', 'processor': 'osint_recon'},
+        follow_redirects=True
+    )
+    assert b"Watch added" in res.data
+    client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
+
+    wait_for_all_checks(client)
+
+    res = client.get(
+        url_for("ui.ui_preview.preview_page", uuid="first"),
+        follow_redirects=True
+    )
+
+    assert b'Target: http://127.0.0.1' in res.data
+    assert b'DNSKEY Records' in res.data
+    wait_for_all_checks(client)
+
+
+    # Now change it to something that doesnt exist
+    uuid = next(iter(live_server.app.config['DATASTORE'].data['watching']))
+    live_server.app.config['DATASTORE'].data['watching'][uuid]['processor'] = "now_missing"
+    client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
+    wait_for_all_checks(client)
+    res = client.get(url_for("watchlist.index"))
+    assert b"Exception: Processor module" in res.data and b'now_missing' in res.data, f'Should register that the plugin is missing for {uuid}'

changedetectionio/tests/test_api.py

@@ -465,7 +465,10 @@ def test_api_watch_PUT_update(client, live_server, measure_memory_usage, datasto
 
     assert res.status_code == 400, "Should get error 400 when we give a field that doesnt exist"
     # Message will come from `flask_expects_json`
-    assert b'Additional properties are not allowed' in res.data
+    # With patternProperties for processor_config_*, the error message format changed slightly
+    assert (b'Additional properties are not allowed' in res.data or
+            b'does not match any of the regexes' in res.data), \
+            "Should reject unknown fields with schema validation error"
 
 
     # Try a XSS URL
@@ -486,6 +489,7 @@ def test_api_import(client, live_server, measure_memory_usage, datastore_path):
 
     api_key = live_server.app.config['DATASTORE'].data['settings']['application'].get('api_access_token')
 
+    # Test 1: Basic import with tag
     res = client.post(
         url_for("import") + "?tag=import-test",
         data='https://website1.com\r\nhttps://website2.com',
@@ -504,6 +508,209 @@ def test_api_import(client, live_server, measure_memory_usage, datastore_path):
     res = client.get(url_for('tags.tags_overview_page'))
     assert b'import-test' in res.data
 
+    # Test 2: Import with watch configuration fields (issue #3845)
+    # Test string field (include_filters), boolean (paused), and processor
+    import urllib.parse
+    params = urllib.parse.urlencode({
+        'tag': 'config-test',
+        'include_filters': 'div.content',
+        'paused': 'true',
+        'processor': 'text_json_diff',
+        'title': 'Imported with Config'
+    })
+
+    res = client.post(
+        url_for("import") + "?" + params,
+        data='https://website3.com',
+        headers={'x-api-key': api_key},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+    assert len(res.json) == 1
+    uuid = res.json[0]
+
+    # Verify the configuration was applied
+    watch = live_server.app.config['DATASTORE'].data['watching'][uuid]
+    assert watch['include_filters'] == ['div.content'], "include_filters should be set as array"
+    assert watch['paused'] == True, "paused should be True"
+    assert watch['processor'] == 'text_json_diff', "processor should be set"
+    assert watch['title'] == 'Imported with Config', "title should be set"
+
+    # Test 3: Import with array field (notification_urls) - using valid Apprise format
+    params = urllib.parse.urlencode({
+        'tag': 'notification-test',
+        'notification_urls': 'mailto://test@example.com,mailto://admin@example.com'
+    })
+
+    res = client.post(
+        url_for("import") + "?" + params,
+        data='https://website4.com',
+        headers={'x-api-key': api_key},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+    uuid = res.json[0]
+    watch = live_server.app.config['DATASTORE'].data['watching'][uuid]
+    assert 'mailto://test@example.com' in watch['notification_urls'], "notification_urls should contain first email"
+    assert 'mailto://admin@example.com' in watch['notification_urls'], "notification_urls should contain second email"
+
+    # Test 4: Import with object field (time_between_check)
+    import json
+    time_config = json.dumps({"hours": 2, "minutes": 30})
+    params = urllib.parse.urlencode({
+        'tag': 'schedule-test',
+        'time_between_check': time_config
+    })
+
+    res = client.post(
+        url_for("import") + "?" + params,
+        data='https://website5.com',
+        headers={'x-api-key': api_key},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+    uuid = res.json[0]
+    watch = live_server.app.config['DATASTORE'].data['watching'][uuid]
+    assert watch['time_between_check']['hours'] == 2, "time_between_check hours should be 2"
+    assert watch['time_between_check']['minutes'] == 30, "time_between_check minutes should be 30"
+
+    # Test 5: Import with invalid processor (should fail)
+    res = client.post(
+        url_for("import") + "?processor=invalid_processor",
+        data='https://website6.com',
+        headers={'x-api-key': api_key},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 400, "Should reject invalid processor"
+    assert b"Invalid processor" in res.data, "Error message should mention invalid processor"
+
+    # Test 6: Import with invalid field (should fail)
+    res = client.post(
+        url_for("import") + "?unknown_field=value",
+        data='https://website7.com',
+        headers={'x-api-key': api_key},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 400, "Should reject unknown field"
+    assert b"Unknown watch configuration parameter" in res.data, "Error message should mention unknown parameter"
+
+
+def test_api_import_small_synchronous(client, live_server, measure_memory_usage, datastore_path):
+    """Test that small imports (< threshold) are processed synchronously"""
+    from changedetectionio.api.Import import IMPORT_SWITCH_TO_BACKGROUND_THRESHOLD
+
+    api_key = live_server.app.config['DATASTORE'].data['settings']['application'].get('api_access_token')
+
+    # Use local test endpoint to avoid network delays
+    test_url_base = url_for('test_endpoint', _external=True)
+
+    # Create URLs: threshold - 1 to stay under limit
+    num_urls = min(5, IMPORT_SWITCH_TO_BACKGROUND_THRESHOLD - 1)  # Use small number for faster test
+    urls = '\n'.join([f'{test_url_base}?id=small-{i}' for i in range(num_urls)])
+
+    # Import small batch
+    res = client.post(
+        url_for("import") + "?tag=small-test",
+        data=urls,
+        headers={'x-api-key': api_key},
+        follow_redirects=True
+    )
+
+    # Should return 200 OK with UUID list (synchronous)
+    assert res.status_code == 200, f"Should return 200 for small imports, got {res.status_code}"
+    assert isinstance(res.json, list), "Response should be a list of UUIDs"
+    assert len(res.json) == num_urls, f"Should return {num_urls} UUIDs, got {len(res.json)}"
+
+    # Verify all watches were created immediately
+    for uuid in res.json:
+        assert uuid in live_server.app.config['DATASTORE'].data['watching'], \
+            f"Watch {uuid} should exist immediately after synchronous import"
+
+    print(f"\n✓ Successfully created {num_urls} watches synchronously")
+
+
+def test_api_import_large_background(client, live_server, measure_memory_usage, datastore_path):
+    """Test that large imports (>= threshold) are processed in background thread"""
+    from changedetectionio.api.Import import IMPORT_SWITCH_TO_BACKGROUND_THRESHOLD
+    import time
+
+    api_key = live_server.app.config['DATASTORE'].data['settings']['application'].get('api_access_token')
+
+    # Use local test endpoint to avoid network delays
+    test_url_base = url_for('test_endpoint', _external=True)
+
+    # Create URLs: threshold + 10 to trigger background processing
+    num_urls = IMPORT_SWITCH_TO_BACKGROUND_THRESHOLD + 10
+    urls = '\n'.join([f'{test_url_base}?id=bulk-{i}' for i in range(num_urls)])
+
+    # Import large batch
+    res = client.post(
+        url_for("import") + "?tag=bulk-test",
+        data=urls,
+        headers={'x-api-key': api_key},
+        follow_redirects=True
+    )
+
+    # Should return 202 Accepted (background processing)
+    assert res.status_code == 202, f"Should return 202 for large imports, got {res.status_code}"
+    assert b"background" in res.data.lower(), "Response should mention background processing"
+
+    # Extract expected count from response
+    response_json = res.json
+    assert 'count' in response_json, "Response should include count"
+    assert response_json['count'] == num_urls, f"Count should be {num_urls}, got {response_json['count']}"
+
+    # Wait for background thread to complete (with timeout)
+    max_wait = 10  # seconds
+    wait_interval = 0.5
+    elapsed = 0
+    watches_created = 0
+
+    while elapsed < max_wait:
+        time.sleep(wait_interval)
+        elapsed += wait_interval
+
+        # Count how many watches have been created
+        watches_created = len([
+            uuid for uuid, watch in live_server.app.config['DATASTORE'].data['watching'].items()
+            if 'id=bulk-' in watch['url']
+        ])
+
+        if watches_created == num_urls:
+            break
+
+    # Verify all watches were created
+    assert watches_created == num_urls, \
+        f"Expected {num_urls} watches to be created, but found {watches_created} after {elapsed}s"
+
+    # Verify watches have correct configuration
+    bulk_watches = [
+        watch for watch in live_server.app.config['DATASTORE'].data['watching'].values()
+        if 'id=bulk-' in watch['url']
+    ]
+
+    assert len(bulk_watches) == num_urls, "All bulk watches should exist"
+
+    # Check that they have the correct tag
+    datastore = live_server.app.config['DATASTORE']
+    # Get UUIDs of bulk watches by filtering the datastore keys
+    bulk_watch_uuids = [
+        uuid for uuid, watch in live_server.app.config['DATASTORE'].data['watching'].items()
+        if 'id=bulk-' in watch['url']
+    ]
+    for watch_uuid in bulk_watch_uuids:
+        tags = datastore.get_all_tags_for_watch(uuid=watch_uuid)
+        tag_names = [t['title'] for t in tags.values()]
+        assert 'bulk-test' in tag_names, f"Watch {watch_uuid} should have 'bulk-test' tag"
+
+    print(f"\n✓ Successfully created {num_urls} watches in background (took {elapsed}s)")
+
+
 def test_api_conflict_UI_password(client, live_server, measure_memory_usage, datastore_path):
 
 

changedetectionio/tests/test_api_notification_urls_validation.py

@@ -0,0 +1,166 @@
+#!/usr/bin/env python3
+
+"""
+Test notification_urls validation in Watch and Tag API endpoints.
+Ensures that invalid AppRise URLs are rejected when setting notification_urls.
+
+Valid AppRise notification URLs use specific protocols like:
+- posts://example.com - POST to HTTP endpoint
+- gets://example.com - GET to HTTP endpoint
+- mailto://user@example.com - Email
+- slack://token/channel - Slack
+- discord://webhook_id/webhook_token - Discord
+- etc.
+
+Invalid notification URLs:
+- https://example.com - Plain HTTPS is NOT a valid AppRise notification protocol
+- ftp://example.com - FTP is NOT a valid AppRise notification protocol
+- Plain URLs without proper AppRise protocol prefix
+"""
+
+from flask import url_for
+import json
+
+
+def test_watch_notification_urls_validation(client, live_server, measure_memory_usage, datastore_path):
+    """Test that Watch PUT/POST endpoints validate notification_urls."""
+    api_key = live_server.app.config['DATASTORE'].data['settings']['application'].get('api_access_token')
+
+    # Test 1: Create a watch with valid notification URLs
+    valid_urls = ["posts://example.com/notify1", "posts://example.com/notify2"]
+    res = client.post(
+        url_for("createwatch"),
+        data=json.dumps({
+            "url": "https://example.com",
+            "notification_urls": valid_urls
+        }),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+    assert res.status_code == 201, "Should accept valid notification URLs on watch creation"
+    watch_uuid = res.json['uuid']
+
+    # Verify the notification URLs were saved
+    res = client.get(
+        url_for("watch", uuid=watch_uuid),
+        headers={'x-api-key': api_key}
+    )
+    assert res.status_code == 200
+    assert set(res.json['notification_urls']) == set(valid_urls), "Valid notification URLs should be saved"
+
+    # Test 2: Try to create a watch with invalid notification URLs (https:// is not valid)
+    invalid_urls = ["https://example.com/webhook"]
+    res = client.post(
+        url_for("createwatch"),
+        data=json.dumps({
+            "url": "https://example.com",
+            "notification_urls": invalid_urls
+        }),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+    assert res.status_code == 400, "Should reject https:// notification URLs (not a valid AppRise protocol)"
+    assert b"is not a valid AppRise URL" in res.data, "Should provide AppRise validation error message"
+
+    # Test 2b: Also test other invalid protocols
+    invalid_urls_ftp = ["ftp://not-apprise-url"]
+    res = client.post(
+        url_for("createwatch"),
+        data=json.dumps({
+            "url": "https://example.com",
+            "notification_urls": invalid_urls_ftp
+        }),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+    assert res.status_code == 400, "Should reject ftp:// notification URLs"
+    assert b"is not a valid AppRise URL" in res.data, "Should provide AppRise validation error message"
+
+    # Test 3: Update watch with valid notification URLs
+    new_valid_urls = ["posts://newserver.com"]
+    res = client.put(
+        url_for("watch", uuid=watch_uuid),
+        data=json.dumps({"notification_urls": new_valid_urls}),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+    assert res.status_code == 200, "Should accept valid notification URLs on watch update"
+
+    # Verify the notification URLs were updated
+    res = client.get(
+        url_for("watch", uuid=watch_uuid),
+        headers={'x-api-key': api_key}
+    )
+    assert res.status_code == 200
+    assert res.json['notification_urls'] == new_valid_urls, "Valid notification URLs should be updated"
+
+    # Test 4: Try to update watch with invalid notification URLs (plain https:// not valid)
+    invalid_https_url = ["https://example.com/webhook"]
+    res = client.put(
+        url_for("watch", uuid=watch_uuid),
+        data=json.dumps({"notification_urls": invalid_https_url}),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+    assert res.status_code == 400, "Should reject https:// notification URLs on watch update"
+    assert b"is not a valid AppRise URL" in res.data, "Should provide AppRise validation error message"
+
+    # Test 5: Update watch with non-list notification_urls (caught by OpenAPI schema validation)
+    res = client.put(
+        url_for("watch", uuid=watch_uuid),
+        data=json.dumps({"notification_urls": "not-a-list"}),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+    assert res.status_code == 400, "Should reject non-list notification_urls"
+    assert b"OpenAPI validation failed" in res.data or b"Request body validation error" in res.data
+
+    # Test 6: Verify original URLs are preserved after failed update
+    res = client.get(
+        url_for("watch", uuid=watch_uuid),
+        headers={'x-api-key': api_key}
+    )
+    assert res.status_code == 200
+    assert res.json['notification_urls'] == new_valid_urls, "URLs should remain unchanged after validation failure"
+
+
+def test_tag_notification_urls_validation(client, live_server, measure_memory_usage, datastore_path):
+    """Test that Tag PUT endpoint validates notification_urls."""
+    from changedetectionio.model import Tag
+
+    api_key = live_server.app.config['DATASTORE'].data['settings']['application'].get('api_access_token')
+    datastore = live_server.app.config['DATASTORE']
+
+    # Create a tag
+    tag_uuid = datastore.add_tag(title="Test Tag")
+    assert tag_uuid is not None
+
+    # Test 1: Update tag with valid notification URLs
+    valid_urls = ["posts://example.com/tag-notify"]
+    res = client.put(
+        url_for("tag", uuid=tag_uuid),
+        data=json.dumps({"notification_urls": valid_urls}),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+    assert res.status_code == 200, "Should accept valid notification URLs on tag update"
+
+    # Verify the notification URLs were saved
+    tag = datastore.data['settings']['application']['tags'][tag_uuid]
+    assert tag['notification_urls'] == valid_urls, "Valid notification URLs should be saved to tag"
+
+    # Test 2: Try to update tag with invalid notification URLs (https:// not valid)
+    invalid_urls = ["https://example.com/webhook"]
+    res = client.put(
+        url_for("tag", uuid=tag_uuid),
+        data=json.dumps({"notification_urls": invalid_urls}),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+    assert res.status_code == 400, "Should reject https:// notification URLs on tag update"
+    assert b"is not a valid AppRise URL" in res.data, "Should provide AppRise validation error message"
+
+    # Test 3: Update tag with non-list notification_urls (caught by OpenAPI schema validation)
+    res = client.put(
+        url_for("tag", uuid=tag_uuid),
+        data=json.dumps({"notification_urls": "not-a-list"}),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+    assert res.status_code == 400, "Should reject non-list notification_urls"
+    assert b"OpenAPI validation failed" in res.data or b"Request body validation error" in res.data
+
+    # Test 4: Verify original URLs are preserved after failed update
+    tag = datastore.data['settings']['application']['tags'][tag_uuid]
+    assert tag['notification_urls'] == valid_urls, "URLs should remain unchanged after validation failure"

changedetectionio/tests/test_api_openapi.py

@@ -80,7 +80,10 @@ def test_openapi_validation_invalid_field_in_request_body(client, live_server, m
     # Should get 400 error due to invalid field (this will be caught by internal validation)
     # Note: This tests the flow where OpenAPI validation passes but internal validation catches it
     assert res.status_code == 400, f"Expected 400 but got {res.status_code}"
-    assert b"Additional properties are not allowed" in res.data, "Should contain validation error about additional properties"
+    # With patternProperties for processor_config_*, the error message format changed slightly
+    assert (b"Additional properties are not allowed" in res.data or
+            b"does not match any of the regexes" in res.data), \
+            "Should contain validation error about additional/invalid properties"
 
 
 def test_openapi_validation_import_wrong_content_type(client, live_server, measure_memory_usage, datastore_path):

changedetectionio/tests/test_api_tags.py

@@ -18,7 +18,7 @@ def test_api_tags_listing(client, live_server, measure_memory_usage, datastore_p
         url_for("tags"),
         headers={'x-api-key': api_key}
     )
-    assert res.text.strip() == "{}", "Should be empty list"
+    assert res.get_data(as_text=True).strip() == "{}", "Should be empty list"
     assert res.status_code == 200
 
     res = client.post(
@@ -36,7 +36,7 @@ def test_api_tags_listing(client, live_server, measure_memory_usage, datastore_p
         headers={'x-api-key': api_key}
     )
     assert res.status_code == 200
-    assert new_tag_uuid in res.text
+    assert new_tag_uuid in res.get_data(as_text=True)
     assert res.json[new_tag_uuid]['title'] == tag_title
     assert res.json[new_tag_uuid]['notification_muted'] == False
 
@@ -118,6 +118,16 @@ def test_api_tags_listing(client, live_server, measure_memory_usage, datastore_p
     assert res.status_code == 200
     assert new_tag_uuid in res.json.get('tags', [])
 
+    # Test that tags are returned when listing ALL watches (issue #3854)
+    res = client.get(
+        url_for("createwatch"),  # GET /api/v1/watch - list all watches
+        headers={'x-api-key': api_key}
+    )
+    assert res.status_code == 200
+    assert watch_uuid in res.json, "Watch should be in the list"
+    assert 'tags' in res.json[watch_uuid], "Tags field should be present in watch list"
+    assert new_tag_uuid in res.json[watch_uuid]['tags'], "Tag UUID should be in tags array"
+
     # Check recheck by tag
     before_check_time = live_server.app.config['DATASTORE'].data['watching'][watch_uuid].get('last_checked')
     time.sleep(1)
@@ -148,7 +158,7 @@ def test_api_tags_listing(client, live_server, measure_memory_usage, datastore_p
         headers={'x-api-key': api_key}
     )
     assert res.status_code == 200
-    assert new_tag_uuid not in res.text
+    assert new_tag_uuid not in res.get_data(as_text=True)
 
     # Verify tag was removed from watch
     res = client.get(

changedetectionio/tests/test_automatic_follow_ldjson_price.py

@@ -2,7 +2,7 @@
 
 import time
 from flask import url_for
-from .util import live_server_setup, extract_UUID_from_client, wait_for_all_checks
+from .util import live_server_setup, extract_UUID_from_client, wait_for_all_checks, delete_all_watches
 import os
 
 
@@ -116,7 +116,7 @@ def test_check_ldjson_price_autodetect(client, live_server, measure_memory_usage
     # And not this cause its not the ld-json
     assert b"So let's see what happens" not in res.data
 
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
     ##########################################################################################
     # And we shouldnt see the offer
@@ -131,7 +131,7 @@ def test_check_ldjson_price_autodetect(client, live_server, measure_memory_usage
     assert b'ldjson-price-track-offer' not in res.data
     
     ##########################################################################################
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 
 def _test_runner_check_bad_format_ignored(live_server, client, has_ldjson_price_data):
@@ -147,7 +147,7 @@ def _test_runner_check_bad_format_ignored(live_server, client, has_ldjson_price_
 
 
     ##########################################################################################
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 
 def test_bad_ldjson_is_correctly_ignored(client, live_server, measure_memory_usage, datastore_path):

changedetectionio/tests/test_backend.py

@@ -414,4 +414,4 @@ def test_plaintext_even_if_xml_content_and_can_apply_filters(client, live_server
     assert b'Abonnementen bijwerken' in res.data
     assert b'<foobar' not in res.data
 
-    res = client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    res = delete_all_watches(client)

changedetectionio/tests/test_backup.py

@@ -53,11 +53,21 @@ def test_backup(client, live_server, measure_memory_usage, datastore_path):
 
     backup = ZipFile(io.BytesIO(res.data))
     l = backup.namelist()
-    uuid4hex = re.compile('^[a-f0-9]{8}-?[a-f0-9]{4}-?4[a-f0-9]{3}-?[89ab][a-f0-9]{3}-?[a-f0-9]{12}.*txt', re.I)
-    newlist = list(filter(uuid4hex.match, l))  # Read Note below
 
+    # Check for UUID-based txt files (history and snapshot)
+    uuid4hex_txt = re.compile('^[a-f0-9]{8}-?[a-f0-9]{4}-?4[a-f0-9]{3}-?[89ab][a-f0-9]{3}-?[a-f0-9]{12}.*txt', re.I)
+    txt_files = list(filter(uuid4hex_txt.match, l))
     # Should be two txt files in the archive (history and the snapshot)
-    assert len(newlist) == 2
+    assert len(txt_files) == 2
+
+    # Check for watch.json files (new format)
+    uuid4hex_json = re.compile('^[a-f0-9]{8}-?[a-f0-9]{4}-?4[a-f0-9]{3}-?[89ab][a-f0-9]{3}-?[a-f0-9]{12}/watch\.json$', re.I)
+    json_files = list(filter(uuid4hex_json.match, l))
+    # Should be one watch.json file in the archive (the imported watch)
+    assert len(json_files) == 1, f"Expected 1 watch.json file, found {len(json_files)}: {json_files}"
+
+    # Check for changedetection.json (settings file)
+    assert 'changedetection.json' in l, "changedetection.json should be in backup"
 
     # Get the latest one
     res = client.get(

changedetectionio/tests/test_basic_socketio.py

@@ -6,7 +6,7 @@ from .util import (
     set_original_response,
     set_modified_response,
     live_server_setup,
-    wait_for_all_checks
+    wait_for_all_checks, delete_all_watches
 )
 from loguru import logger
 
@@ -104,7 +104,7 @@ def run_socketio_watch_update_test(client, live_server, password_mode="", datast
     assert watch.has_unviewed, "The watch was not marked as unviewed after content change"
 
     # Clean up
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 def test_everything(live_server, client, measure_memory_usage, datastore_path):
 

changedetectionio/tests/test_commit_persistence.py

@@ -0,0 +1,661 @@
+#!/usr/bin/env python3
+"""
+Tests for immediate commit-based persistence system.
+
+Tests cover:
+- Watch.commit() persistence to disk
+- Concurrent commit safety (race conditions)
+- Processor config separation
+- Data loss prevention (settings, tags, watch modifications)
+"""
+
+import json
+import os
+import threading
+import time
+from flask import url_for
+from .util import wait_for_all_checks
+
+
+# ==============================================================================
+# 2. Commit() Persistence Tests
+# ==============================================================================
+
+def test_watch_commit_persists_to_disk(client, live_server):
+    """Test that watch.commit() actually writes to watch.json immediately"""
+    datastore = client.application.config.get('DATASTORE')
+
+    # Create a watch
+    uuid = datastore.add_watch(url='http://example.com', extras={'title': 'Original Title'})
+    watch = datastore.data['watching'][uuid]
+
+    # Modify and commit
+    watch['title'] = 'Modified Title'
+    watch['paused'] = True
+    watch.commit()
+
+    # Read directly from disk (bypass datastore cache)
+    watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+    assert os.path.exists(watch_json_path), "watch.json should exist on disk"
+
+    with open(watch_json_path, 'r') as f:
+        disk_data = json.load(f)
+
+    assert disk_data['title'] == 'Modified Title', "Title should be persisted to disk"
+    assert disk_data['paused'] == True, "Paused state should be persisted to disk"
+    assert disk_data['uuid'] == uuid, "UUID should match"
+
+
+def test_watch_commit_survives_reload(client, live_server):
+    """Test that committed changes survive datastore reload"""
+    from changedetectionio.store import ChangeDetectionStore
+
+    datastore = client.application.config.get('DATASTORE')
+    datastore_path = datastore.datastore_path
+
+    # Create and modify a watch
+    uuid = datastore.add_watch(url='http://example.com', extras={'title': 'Test Watch'})
+    watch = datastore.data['watching'][uuid]
+    watch['title'] = 'Persisted Title'
+    watch['paused'] = True
+    watch['tags'] = ['tag-1', 'tag-2']
+    watch.commit()
+
+    # Simulate app restart - create new datastore instance
+    datastore2 = ChangeDetectionStore(datastore_path=datastore_path)
+    datastore2.reload_state(
+        datastore_path=datastore_path,
+        include_default_watches=False,
+        version_tag='test'
+    )
+
+    # Check data survived
+    assert uuid in datastore2.data['watching'], "Watch should exist after reload"
+    reloaded_watch = datastore2.data['watching'][uuid]
+    assert reloaded_watch['title'] == 'Persisted Title', "Title should survive reload"
+    assert reloaded_watch['paused'] == True, "Paused state should survive reload"
+    assert reloaded_watch['tags'] == ['tag-1', 'tag-2'], "Tags should survive reload"
+
+
+def test_watch_commit_atomic_on_crash(client, live_server):
+    """Test that atomic writes prevent corruption (temp file pattern)"""
+    datastore = client.application.config.get('DATASTORE')
+
+    uuid = datastore.add_watch(url='http://example.com', extras={'title': 'Original'})
+    watch = datastore.data['watching'][uuid]
+
+    # First successful commit
+    watch['title'] = 'First Save'
+    watch.commit()
+
+    # Verify watch.json exists and is valid
+    watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)  # Should not raise JSONDecodeError
+        assert data['title'] == 'First Save'
+
+    # Second commit - even if interrupted, original file should be intact
+    # (atomic write uses temp file + rename, so original is never corrupted)
+    watch['title'] = 'Second Save'
+    watch.commit()
+
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)
+        assert data['title'] == 'Second Save'
+
+
+def test_multiple_watches_commit_independently(client, live_server):
+    """Test that committing one watch doesn't affect others"""
+    datastore = client.application.config.get('DATASTORE')
+
+    # Create multiple watches
+    uuid1 = datastore.add_watch(url='http://example1.com', extras={'title': 'Watch 1'})
+    uuid2 = datastore.add_watch(url='http://example2.com', extras={'title': 'Watch 2'})
+    uuid3 = datastore.add_watch(url='http://example3.com', extras={'title': 'Watch 3'})
+
+    watch1 = datastore.data['watching'][uuid1]
+    watch2 = datastore.data['watching'][uuid2]
+    watch3 = datastore.data['watching'][uuid3]
+
+    # Modify and commit only watch2
+    watch2['title'] = 'Modified Watch 2'
+    watch2['paused'] = True
+    watch2.commit()
+
+    # Read all from disk
+    def read_watch_json(uuid):
+        watch = datastore.data['watching'][uuid]
+        path = os.path.join(watch.data_dir, 'watch.json')
+        with open(path, 'r') as f:
+            return json.load(f)
+
+    data1 = read_watch_json(uuid1)
+    data2 = read_watch_json(uuid2)
+    data3 = read_watch_json(uuid3)
+
+    # Only watch2 should have changes
+    assert data1['title'] == 'Watch 1', "Watch 1 should be unchanged"
+    assert data1['paused'] == False, "Watch 1 should not be paused"
+
+    assert data2['title'] == 'Modified Watch 2', "Watch 2 should be modified"
+    assert data2['paused'] == True, "Watch 2 should be paused"
+
+    assert data3['title'] == 'Watch 3', "Watch 3 should be unchanged"
+    assert data3['paused'] == False, "Watch 3 should not be paused"
+
+
+# ==============================================================================
+# 3. Concurrency/Race Condition Tests
+# ==============================================================================
+
+def test_concurrent_watch_commits_dont_corrupt(client, live_server):
+    """Test that simultaneous commits to same watch don't corrupt JSON"""
+    datastore = client.application.config.get('DATASTORE')
+
+    uuid = datastore.add_watch(url='http://example.com', extras={'title': 'Test'})
+    watch = datastore.data['watching'][uuid]
+
+    errors = []
+
+    def modify_and_commit(field, value):
+        try:
+            watch[field] = value
+            watch.commit()
+        except Exception as e:
+            errors.append(e)
+
+    # Run 10 concurrent commits
+    threads = []
+    for i in range(10):
+        t = threading.Thread(target=modify_and_commit, args=('title', f'Title {i}'))
+        threads.append(t)
+        t.start()
+
+    for t in threads:
+        t.join()
+
+    # Should not have any errors
+    assert len(errors) == 0, f"Expected no errors, got: {errors}"
+
+    # JSON file should still be valid (not corrupted)
+    watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)  # Should not raise JSONDecodeError
+        assert data['uuid'] == uuid, "UUID should still be correct"
+        assert 'Title' in data['title'], "Title should contain 'Title'"
+
+
+def test_concurrent_modifications_during_commit(client, live_server):
+    """Test that modifying watch during commit doesn't cause RuntimeError"""
+    datastore = client.application.config.get('DATASTORE')
+
+    uuid = datastore.add_watch(url='http://example.com', extras={'title': 'Test'})
+    watch = datastore.data['watching'][uuid]
+
+    errors = []
+    stop_flag = threading.Event()
+
+    def keep_modifying():
+        """Continuously modify watch"""
+        try:
+            i = 0
+            while not stop_flag.is_set():
+                watch['title'] = f'Title {i}'
+                watch['paused'] = i % 2 == 0
+                i += 1
+                time.sleep(0.001)
+        except Exception as e:
+            errors.append(('modifier', e))
+
+    def keep_committing():
+        """Continuously commit watch"""
+        try:
+            for _ in range(20):
+                watch.commit()
+                time.sleep(0.005)
+        except Exception as e:
+            errors.append(('committer', e))
+
+    # Start concurrent modification and commits
+    modifier = threading.Thread(target=keep_modifying)
+    committer = threading.Thread(target=keep_committing)
+
+    modifier.start()
+    committer.start()
+
+    committer.join()
+    stop_flag.set()
+    modifier.join()
+
+    # Should not have RuntimeError from dict changing during iteration
+    runtime_errors = [e for source, e in errors if isinstance(e, RuntimeError)]
+    assert len(runtime_errors) == 0, f"Should not have RuntimeError, got: {runtime_errors}"
+
+
+def test_datastore_lock_protects_commit_snapshot(client, live_server):
+    """Test that datastore.lock prevents race conditions during deepcopy"""
+    datastore = client.application.config.get('DATASTORE')
+
+    uuid = datastore.add_watch(url='http://example.com', extras={'title': 'Test'})
+    watch = datastore.data['watching'][uuid]
+
+    # Add some complex nested data
+    watch['browser_steps'] = [
+        {'operation': 'click', 'selector': '#foo'},
+        {'operation': 'wait', 'seconds': 5}
+    ]
+
+    errors = []
+    commits_succeeded = [0]
+
+    def rapid_commits():
+        try:
+            for i in range(50):
+                watch['title'] = f'Title {i}'
+                watch.commit()
+                commits_succeeded[0] += 1
+                time.sleep(0.001)
+        except Exception as e:
+            errors.append(e)
+
+    # Multiple threads doing rapid commits
+    threads = [threading.Thread(target=rapid_commits) for _ in range(3)]
+
+    for t in threads:
+        t.start()
+    for t in threads:
+        t.join()
+
+    assert len(errors) == 0, f"Expected no errors, got: {errors}"
+    assert commits_succeeded[0] == 150, f"Expected 150 commits, got {commits_succeeded[0]}"
+
+    # Final JSON should be valid
+    watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)
+        assert data['uuid'] == uuid
+
+
+# ==============================================================================
+# 4. Processor Config Separation Tests
+# ==============================================================================
+
+def test_processor_config_never_in_watch_json(client, live_server):
+    """Test that processor_config_* fields are filtered out of watch.json"""
+    datastore = client.application.config.get('DATASTORE')
+
+    uuid = datastore.add_watch(
+        url='http://example.com',
+        extras={
+            'title': 'Test Watch',
+            'processor': 'restock_diff'
+        }
+    )
+
+    watch = datastore.data['watching'][uuid]
+
+    # Try to set processor config fields (these should be filtered during commit)
+    watch['processor_config_price_threshold'] = 10.0
+    watch['processor_config_some_setting'] = 'value'
+    watch['processor_config_another'] = {'nested': 'data'}
+    watch.commit()
+
+    # Read watch.json from disk
+    watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)
+
+    # Verify processor_config_* fields are NOT in watch.json
+    for key in data.keys():
+        assert not key.startswith('processor_config_'), \
+            f"Found {key} in watch.json - processor configs should be in separate file!"
+
+    # Normal fields should still be there
+    assert data['title'] == 'Test Watch'
+    assert data['processor'] == 'restock_diff'
+
+
+def test_api_post_saves_processor_config_separately(client, live_server):
+    """Test that API POST saves processor configs to {processor}.json"""
+    import json
+    from changedetectionio.processors import extract_processor_config_from_form_data
+
+    # Get API key
+    api_key = live_server.app.config['DATASTORE'].data['settings']['application'].get('api_access_token')
+
+    # Create watch via API with processor config
+    response = client.post(
+        url_for("createwatch"),
+        data=json.dumps({
+            'url': 'http://example.com',
+            'processor': 'restock_diff',
+            'processor_config_price_threshold': 10.0,
+            'processor_config_in_stock_only': True
+        }),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+
+    assert response.status_code in (200, 201), f"Expected 200/201, got {response.status_code}"
+    uuid = response.json.get('uuid')
+    assert uuid, "Should return UUID"
+
+    datastore = client.application.config.get('DATASTORE')
+    watch = datastore.data['watching'][uuid]
+
+    # Check that processor config file exists
+    processor_config_path = os.path.join(watch.data_dir, 'restock_diff.json')
+    assert os.path.exists(processor_config_path), "Processor config file should exist"
+
+    with open(processor_config_path, 'r') as f:
+        config = json.load(f)
+
+    # Verify fields are saved WITHOUT processor_config_ prefix
+    assert config.get('price_threshold') == 10.0, "Should have price_threshold (no prefix)"
+    assert config.get('in_stock_only') == True, "Should have in_stock_only (no prefix)"
+    assert 'processor_config_price_threshold' not in config, "Should NOT have prefixed keys"
+
+
+def test_api_put_saves_processor_config_separately(client, live_server):
+    """Test that API PUT updates processor configs in {processor}.json"""
+    import json
+    datastore = client.application.config.get('DATASTORE')
+
+    # Get API key
+    api_key = live_server.app.config['DATASTORE'].data['settings']['application'].get('api_access_token')
+
+    # Create watch
+    uuid = datastore.add_watch(
+        url='http://example.com',
+        extras={'processor': 'restock_diff'}
+    )
+
+    # Update via API with processor config
+    response = client.put(
+        url_for("watch", uuid=uuid),
+        data=json.dumps({
+            'processor_config_price_threshold': 15.0,
+            'processor_config_min_stock': 5
+        }),
+        headers={'content-type': 'application/json', 'x-api-key': api_key}
+    )
+
+    # PUT might return different status codes, 200 or 204 are both OK
+    assert response.status_code in (200, 204), f"Expected 200/204, got {response.status_code}: {response.data}"
+
+    watch = datastore.data['watching'][uuid]
+
+    # Check processor config file
+    processor_config_path = os.path.join(watch.data_dir, 'restock_diff.json')
+    assert os.path.exists(processor_config_path), "Processor config file should exist"
+
+    with open(processor_config_path, 'r') as f:
+        config = json.load(f)
+
+    assert config.get('price_threshold') == 15.0, "Should have updated price_threshold"
+    assert config.get('min_stock') == 5, "Should have min_stock"
+
+
+def test_ui_edit_saves_processor_config_separately(client, live_server):
+    """Test that processor_config_* fields never appear in watch.json (even from UI)"""
+    datastore = client.application.config.get('DATASTORE')
+
+    # Create watch
+    uuid = datastore.add_watch(
+        url='http://example.com',
+        extras={'processor': 'text_json_diff', 'title': 'Test'}
+    )
+
+    watch = datastore.data['watching'][uuid]
+
+    # Simulate someone accidentally trying to set processor_config fields directly
+    watch['processor_config_should_not_save'] = 'test_value'
+    watch['processor_config_another_field'] = 123
+    watch['normal_field'] = 'this_should_save'
+    watch.commit()
+
+    # Check watch.json has NO processor_config_* fields (main point of this test)
+    watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+    with open(watch_json_path, 'r') as f:
+        watch_data = json.load(f)
+
+    for key in watch_data.keys():
+        assert not key.startswith('processor_config_'), \
+            f"Found {key} in watch.json - processor configs should be filtered during commit"
+
+    # Verify normal fields still save
+    assert watch_data['normal_field'] == 'this_should_save', "Normal fields should save"
+    assert watch_data['title'] == 'Test', "Original fields should still be there"
+
+
+def test_browser_steps_normalized_to_empty_list(client, live_server):
+    """Test that meaningless browser_steps are normalized to [] during commit"""
+    datastore = client.application.config.get('DATASTORE')
+
+    uuid = datastore.add_watch(url='http://example.com')
+    watch = datastore.data['watching'][uuid]
+
+    # Set browser_steps to meaningless values
+    watch['browser_steps'] = [
+        {'operation': 'Choose one', 'selector': ''},
+        {'operation': 'Goto site', 'selector': ''},
+        {'operation': '', 'selector': '#foo'}
+    ]
+    watch.commit()
+
+    # Read from disk
+    watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)
+
+    # Should be normalized to empty list
+    assert data['browser_steps'] == [], "Meaningless browser_steps should be normalized to []"
+
+
+# ==============================================================================
+# 5. Data Loss Prevention Tests
+# ==============================================================================
+
+def test_settings_persist_after_update(client, live_server):
+    """Test that settings updates are committed and survive restart"""
+    from changedetectionio.store import ChangeDetectionStore
+
+    datastore = client.application.config.get('DATASTORE')
+    datastore_path = datastore.datastore_path
+
+    # Update settings directly (bypass form validation issues)
+    datastore.data['settings']['application']['empty_pages_are_a_change'] = True
+    datastore.data['settings']['application']['fetch_backend'] = 'html_requests'
+    datastore.data['settings']['requests']['time_between_check']['minutes'] = 120
+    datastore.commit()
+
+    # Simulate restart
+    datastore2 = ChangeDetectionStore(datastore_path=datastore_path)
+    datastore2.reload_state(
+        datastore_path=datastore_path,
+        include_default_watches=False,
+        version_tag='test'
+    )
+
+    # Verify settings survived
+    assert datastore2.data['settings']['application']['empty_pages_are_a_change'] == True, "empty_pages_are_a_change should persist"
+    assert datastore2.data['settings']['application']['fetch_backend'] == 'html_requests', "fetch_backend should persist"
+    assert datastore2.data['settings']['requests']['time_between_check']['minutes'] == 120, "time_between_check should persist"
+
+
+def test_tag_mute_persists(client, live_server):
+    """Test that tag mute/unmute operations persist"""
+    from changedetectionio.store import ChangeDetectionStore
+
+    datastore = client.application.config.get('DATASTORE')
+    datastore_path = datastore.datastore_path
+
+    # Add a tag
+    tag_uuid = datastore.add_tag('Test Tag')
+
+    # Mute the tag
+    response = client.get(url_for("tags.mute", uuid=tag_uuid))
+    assert response.status_code == 302  # Redirect
+
+    # Verify muted in memory
+    assert datastore.data['settings']['application']['tags'][tag_uuid]['notification_muted'] == True
+
+    # Simulate restart
+    datastore2 = ChangeDetectionStore(datastore_path=datastore_path)
+    datastore2.reload_state(
+        datastore_path=datastore_path,
+        include_default_watches=False,
+        version_tag='test'
+    )
+
+    # Verify mute state survived
+    assert tag_uuid in datastore2.data['settings']['application']['tags']
+    assert datastore2.data['settings']['application']['tags'][tag_uuid]['notification_muted'] == True
+
+
+def test_tag_delete_removes_from_watches(client, live_server):
+    """Test that deleting a tag removes it from all watches"""
+    datastore = client.application.config.get('DATASTORE')
+
+    # Create a tag
+    tag_uuid = datastore.add_tag('Test Tag')
+
+    # Create watches with this tag
+    uuid1 = datastore.add_watch(url='http://example1.com')
+    uuid2 = datastore.add_watch(url='http://example2.com')
+    uuid3 = datastore.add_watch(url='http://example3.com')
+
+    watch1 = datastore.data['watching'][uuid1]
+    watch2 = datastore.data['watching'][uuid2]
+    watch3 = datastore.data['watching'][uuid3]
+
+    watch1['tags'] = [tag_uuid]
+    watch1.commit()
+    watch2['tags'] = [tag_uuid, 'other-tag']
+    watch2.commit()
+    # watch3 has no tags
+
+    # Delete the tag
+    response = client.get(url_for("tags.delete", uuid=tag_uuid))
+    assert response.status_code == 302
+
+    # Wait for background thread to complete
+    time.sleep(1)
+
+    # Tag should be removed from settings
+    assert tag_uuid not in datastore.data['settings']['application']['tags']
+
+    # Tag should be removed from watches and persisted
+    def check_watch_tags(uuid):
+        watch = datastore.data['watching'][uuid]
+        watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+        with open(watch_json_path, 'r') as f:
+            return json.load(f)['tags']
+
+    assert tag_uuid not in check_watch_tags(uuid1), "Tag should be removed from watch1"
+    assert tag_uuid not in check_watch_tags(uuid2), "Tag should be removed from watch2"
+    assert 'other-tag' in check_watch_tags(uuid2), "Other tags should remain in watch2"
+    assert check_watch_tags(uuid3) == [], "Watch3 should still have empty tags"
+
+
+def test_watch_pause_unpause_persists(client, live_server):
+    """Test that pause/unpause operations commit and persist"""
+    datastore = client.application.config.get('DATASTORE')
+
+    # Get API key
+    api_key = live_server.app.config['DATASTORE'].data['settings']['application'].get('api_access_token')
+
+    uuid = datastore.add_watch(url='http://example.com')
+    watch = datastore.data['watching'][uuid]
+
+    # Pause via API
+    response = client.get(url_for("watch", uuid=uuid, paused='paused'), headers={'x-api-key': api_key})
+    assert response.status_code == 200
+
+    # Check persisted to disk
+    watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)
+    assert data['paused'] == True, "Pause should be persisted"
+
+    # Unpause
+    response = client.get(url_for("watch", uuid=uuid, paused='unpaused'), headers={'x-api-key': api_key})
+    assert response.status_code == 200
+
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)
+    assert data['paused'] == False, "Unpause should be persisted"
+
+
+def test_watch_mute_unmute_persists(client, live_server):
+    """Test that mute/unmute operations commit and persist"""
+    datastore = client.application.config.get('DATASTORE')
+
+    # Get API key
+    api_key = live_server.app.config['DATASTORE'].data['settings']['application'].get('api_access_token')
+
+    uuid = datastore.add_watch(url='http://example.com')
+    watch = datastore.data['watching'][uuid]
+
+    # Mute via API
+    response = client.get(url_for("watch", uuid=uuid, muted='muted'), headers={'x-api-key': api_key})
+    assert response.status_code == 200
+
+    # Check persisted to disk
+    watch_json_path = os.path.join(watch.data_dir, 'watch.json')
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)
+    assert data['notification_muted'] == True, "Mute should be persisted"
+
+    # Unmute
+    response = client.get(url_for("watch", uuid=uuid, muted='unmuted'), headers={'x-api-key': api_key})
+    assert response.status_code == 200
+
+    with open(watch_json_path, 'r') as f:
+        data = json.load(f)
+    assert data['notification_muted'] == False, "Unmute should be persisted"
+
+
+def test_ui_watch_edit_persists_all_fields(client, live_server):
+    """Test that UI watch edit form persists all modified fields"""
+    from changedetectionio.store import ChangeDetectionStore
+
+    datastore = client.application.config.get('DATASTORE')
+    datastore_path = datastore.datastore_path
+
+    # Create watch
+    uuid = datastore.add_watch(url='http://example.com')
+
+    # Edit via UI with multiple field changes
+    response = client.post(
+        url_for("ui.ui_edit.edit_page", uuid=uuid),
+        data={
+            'url': 'http://updated-example.com',
+            'title': 'Updated Watch Title',
+            'time_between_check-hours': '2',
+            'time_between_check-minutes': '30',
+            'include_filters': '#content',
+            'fetch_backend': 'html_requests',
+            'method': 'POST',
+            'ignore_text': 'Advertisement\nTracking'
+        },
+        follow_redirects=True
+    )
+
+    assert b"Updated watch" in response.data or b"Saved" in response.data
+
+    # Simulate restart
+    datastore2 = ChangeDetectionStore(datastore_path=datastore_path)
+    datastore2.reload_state(
+        datastore_path=datastore_path,
+        include_default_watches=False,
+        version_tag='test'
+    )
+
+    # Verify all fields survived
+    watch = datastore2.data['watching'][uuid]
+    assert watch['url'] == 'http://updated-example.com'
+    assert watch['title'] == 'Updated Watch Title'
+    assert watch['time_between_check']['hours'] == 2
+    assert watch['time_between_check']['minutes'] == 30
+    assert watch['fetch_backend'] == 'html_requests'
+    assert watch['method'] == 'POST'

changedetectionio/tests/test_conditions.py

@@ -69,7 +69,7 @@ def test_conditions_with_text_and_number(client, live_server, measure_memory_usa
     # 1. The page filtered text must contain "5" (first digit of value)
     # 2. The extracted number should be >= 20 and <= 100
     res = client.post(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuid),
         data={
             "url": test_url,
             "fetch_backend": "html_requests",
@@ -110,25 +110,20 @@ def test_conditions_with_text_and_number(client, live_server, measure_memory_usa
 
     wait_for_all_checks(client)
     client.get(url_for("ui.mark_all_viewed"), follow_redirects=True)
-    time.sleep(0.2)
-
-    wait_for_all_checks(client)
+    time.sleep(1)
 
     # Case 1
     set_number_in_range_response(datastore_path=datastore_path, number="70.5")
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
     wait_for_all_checks(client)
 
-    time.sleep(2)
     # 75 is > 20 and < 100 and contains "5"
     res = client.get(url_for("watchlist.index"))
     assert b'has-unread-changes' in res.data
 
-
     # Case 2: Change with one condition violated
     # Number out of range (150) but contains '5'
     client.get(url_for("ui.mark_all_viewed"), follow_redirects=True)
-    time.sleep(0.2)
 
     set_number_out_of_range_response(datastore_path=datastore_path, number="150.5")
 
@@ -154,7 +149,6 @@ def test_condition_validate_rule_row(client, live_server, measure_memory_usage,
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
     wait_for_all_checks(client)
 
-    uuid = next(iter(live_server.app.config['DATASTORE'].data['watching']))
 
     # the front end submits the current form state which should override the watch in a temporary copy
     res = client.post(
@@ -195,12 +189,8 @@ def test_condition_validate_rule_row(client, live_server, measure_memory_usage,
     )
     assert res.status_code == 200
     assert b'false' in res.data
-    # cleanup for the next
-    client.get(
-        url_for("ui.form_delete", uuid="all"),
-        follow_redirects=True
-    )
 
+    delete_all_watches(client)
 
 
 # If there was only a change in the whitespacing, then we shouldnt have a change detected
@@ -230,17 +220,12 @@ def test_wordcount_conditions_plugin(client, live_server, measure_memory_usage,
 
     # Check it saved
     res = client.get(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuid),
     )
 
     # Assert the word count is counted correctly
     assert b'<td>13</td>' in res.data
-
-    # cleanup for the next
-    client.get(
-        url_for("ui.form_delete", uuid="all"),
-        follow_redirects=True
-    )
+    delete_all_watches(client)
 
 # If there was only a change in the whitespacing, then we shouldnt have a change detected
 def test_lev_conditions_plugin(client, live_server, measure_memory_usage, datastore_path):

changedetectionio/tests/test_errorhandling.py

@@ -64,6 +64,7 @@ def test_DNS_errors(client, live_server, measure_memory_usage, datastore_path):
         follow_redirects=True
     )
     assert b"1 Imported" in res.data
+    res = client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
 
     # Give the thread time to pick it up
     wait_for_all_checks(client)
@@ -79,7 +80,7 @@ def test_DNS_errors(client, live_server, measure_memory_usage, datastore_path):
     )
     assert found_name_resolution_error
     # Should always record that we tried
-    assert bytes("just now".encode('utf-8')) in res.data
+    assert "just now".encode('utf-8') in res.data or 'seconds ago'.encode('utf-8') in res.data
     delete_all_watches(client)
 
 # Re 1513

changedetectionio/tests/test_filter_failure_notification.py

@@ -1,7 +1,7 @@
 import os
 import time
 from flask import url_for
-from .util import set_original_response,  wait_for_all_checks, wait_for_notification_endpoint_output
+from .util import set_original_response, wait_for_all_checks, wait_for_notification_endpoint_output, delete_all_watches
 from ..notification import valid_notification_formats
 
 
@@ -118,8 +118,10 @@ def run_filter_test(client, live_server, content_filter, app_notification_format
         res = client.get(url_for("watchlist.index"))
         assert b'Warning, no filters were found' in res.data
         assert not os.path.isfile(notification_file)
-        time.sleep(1)
+        time.sleep(2)
+        wait_for_all_checks(client)
 
+    wait_for_all_checks(client)
     assert live_server.app.config['DATASTORE'].data['watching'][uuid]['consecutive_filter_failures'] == 5
 
     time.sleep(2)
@@ -178,6 +180,7 @@ def run_filter_test(client, live_server, content_filter, app_notification_format
         follow_redirects=True
     )
     os.unlink(notification_file)
+    delete_all_watches(client)
 
 
 def test_check_include_filters_failure_notification(client, live_server, measure_memory_usage, datastore_path):
@@ -185,10 +188,12 @@ def test_check_include_filters_failure_notification(client, live_server, measure
     run_filter_test(client=client, live_server=live_server, content_filter='#nope-doesnt-exist', app_notification_format=valid_notification_formats.get('htmlcolor'), datastore_path=datastore_path)
     # Check markup send conversion didnt affect plaintext preference
     run_filter_test(client=client, live_server=live_server, content_filter='#nope-doesnt-exist', app_notification_format=valid_notification_formats.get('text'), datastore_path=datastore_path)
+    delete_all_watches(client)
 
 def test_check_xpath_filter_failure_notification(client, live_server, measure_memory_usage, datastore_path):
     #   #  live_server_setup(live_server) # Setup on conftest per function
     run_filter_test(client=client, live_server=live_server, content_filter='//*[@id="nope-doesnt-exist"]', app_notification_format=valid_notification_formats.get('htmlcolor'), datastore_path=datastore_path)
+    delete_all_watches(client)
 
 # Test that notification is never sent
 
@@ -197,3 +202,4 @@ def test_basic_markup_from_text(client, live_server, measure_memory_usage, datas
     from ..notification.handler import markup_text_links_to_html
     x = markup_text_links_to_html("hello https://google.com")
     assert 'a href' in x
+    delete_all_watches(client)

changedetectionio/tests/test_group.py

@@ -5,6 +5,8 @@ from flask import url_for
 from .util import live_server_setup, wait_for_all_checks, extract_rss_token_from_UI, get_UUID_for_tag_name, extract_UUID_from_client, delete_all_watches
 import os
 
+from ..store import ChangeDetectionStore
+
 
 # def test_setup(client, live_server, measure_memory_usage, datastore_path):
    #  live_server_setup(live_server) # Setup on conftest per function
@@ -166,7 +168,8 @@ def test_tag_add_in_ui(client, live_server, measure_memory_usage, datastore_path
     delete_all_watches(client)
 
 def test_group_tag_notification(client, live_server, measure_memory_usage, datastore_path):
-    
+    delete_all_watches(client)
+
     set_original_response(datastore_path=datastore_path)
 
     test_url = url_for('test_endpoint', _external=True)
@@ -473,3 +476,143 @@ the {test} appeared before. {test in res.data[:n]=}
         n += t_index + len(test)
 
     delete_all_watches(client)
+
+
+def test_tag_json_persistence(client, live_server, measure_memory_usage, datastore_path):
+    """
+    Test that tags are saved to individual tag.json files and loaded correctly.
+
+    This test verifies the update_27 tag storage refactoring:
+    - Tags are saved to {uuid}/tag.json files
+    - Tags persist across datastore restarts
+    - Tag edits write to tag.json
+    - Tag deletion removes tag.json file
+    """
+    import json
+
+    datastore = client.application.config.get('DATASTORE')
+
+    # 1. Create a tag
+    res = client.post(
+        url_for("tags.form_tag_add"),
+        data={"name": "persistence-test-tag"},
+        follow_redirects=True
+    )
+    assert b"Tag added" in res.data
+
+    tag_uuid = get_UUID_for_tag_name(client, name="persistence-test-tag")
+    assert tag_uuid, "Tag UUID should exist"
+
+    # 2. Verify tag.json file was created
+    tag_json_path = os.path.join(datastore_path, tag_uuid, "tag.json")
+    assert os.path.exists(tag_json_path), f"tag.json should exist at {tag_json_path}"
+
+    # 3. Verify tag.json contains correct data
+    with open(tag_json_path, 'r') as f:
+        tag_data = json.load(f)
+    assert tag_data['title'] == 'persistence-test-tag'
+    assert tag_data['uuid'] == tag_uuid
+    assert 'date_created' in tag_data
+
+    # 4. Edit the tag
+    res = client.post(
+        url_for("tags.form_tag_edit_submit", uuid=tag_uuid),
+        data={
+            "name": "persistence-test-tag",
+            "notification_muted": True,
+            "include_filters": '#test-filter'
+        },
+        follow_redirects=True
+    )
+    assert b"Updated" in res.data
+
+    # 5. Verify tag.json was updated
+    with open(tag_json_path, 'r') as f:
+        tag_data = json.load(f)
+    assert tag_data['notification_muted'] == True
+    assert '#test-filter' in tag_data.get('include_filters', [])
+
+    # 5a. Verify tag is NOT in changedetection.json (tags should be in tag.json only)
+    changedetection_json_path = os.path.join(datastore_path, "changedetection.json")
+    with open(changedetection_json_path, 'r') as f:
+        settings_data = json.load(f)
+    # Tags dict should be empty in settings (all tags are in individual files)
+    assert settings_data['settings']['application']['tags'] == {}, \
+        "Tags should NOT be saved to changedetection.json (should be empty dict)"
+
+    # 6. Simulate restart - reload datastore
+    datastore2 = ChangeDetectionStore(datastore_path=datastore_path, include_default_watches=False, version_tag='test')
+
+    # 7. Verify tag was loaded from tag.json
+    assert tag_uuid in datastore2.data['settings']['application']['tags']
+    loaded_tag = datastore2.data['settings']['application']['tags'][tag_uuid]
+    assert loaded_tag['title'] == 'persistence-test-tag'
+    assert loaded_tag['notification_muted'] == True
+    assert '#test-filter' in loaded_tag.get('include_filters', [])
+
+    # 8. Delete the tag via API
+    res = client.get(url_for("tags.delete", uuid=tag_uuid), follow_redirects=True)
+    assert b"Tag deleted" in res.data
+
+    # 9. Verify tag.json file was deleted
+    assert not os.path.exists(tag_json_path), f"tag.json should be deleted at {tag_json_path}"
+
+    # 10. Verify tag is removed from settings
+    assert tag_uuid not in datastore.data['settings']['application']['tags']
+
+    delete_all_watches(client)
+
+
+def test_tag_json_migration_update_27(client, live_server, measure_memory_usage, datastore_path):
+    """
+    Test that update_27 migration correctly moves tags to individual files.
+
+    This simulates a pre-update_27 datastore and verifies migration works.
+    """
+    import json
+
+    # 1. Create multiple tags
+    tag_names = ['migration-tag-1', 'migration-tag-2', 'migration-tag-3']
+    tag_uuids = []
+
+    for tag_name in tag_names:
+        res = client.post(
+            url_for("tags.form_tag_add"),
+            data={"name": tag_name},
+            follow_redirects=True
+        )
+        assert b"Tag added" in res.data
+        tag_uuid = get_UUID_for_tag_name(client, name=tag_name)
+        tag_uuids.append(tag_uuid)
+
+    # 2. Verify all tag.json files exist (update_27 already ran during add_tag)
+    for tag_uuid in tag_uuids:
+        tag_json_path = os.path.join(datastore_path, tag_uuid, "tag.json")
+        assert os.path.exists(tag_json_path), f"tag.json should exist for {tag_uuid}"
+
+    # 2a. Verify tags are NOT in changedetection.json
+    changedetection_json_path = os.path.join(datastore_path, "changedetection.json")
+    with open(changedetection_json_path, 'r') as f:
+        settings_data = json.load(f)
+    assert settings_data['settings']['application']['tags'] == {}, \
+        "Tags should NOT be in changedetection.json after migration"
+
+    # 3. Simulate restart
+    datastore2 = ChangeDetectionStore(datastore_path=datastore_path, include_default_watches=False, version_tag='test')
+
+    # 4. Verify all tags loaded from tag.json files
+    for idx, tag_uuid in enumerate(tag_uuids):
+        assert tag_uuid in datastore2.data['settings']['application']['tags']
+        loaded_tag = datastore2.data['settings']['application']['tags'][tag_uuid]
+        assert loaded_tag['title'] == tag_names[idx]
+
+    # Cleanup
+    res = client.get(url_for("tags.delete_all"), follow_redirects=True)
+    assert b'All tags deleted' in res.data
+
+    # Verify all tag.json files were deleted
+    for tag_uuid in tag_uuids:
+        tag_json_path = os.path.join(datastore_path, tag_uuid, "tag.json")
+        assert not os.path.exists(tag_json_path), f"tag.json should be deleted for {tag_uuid}"
+
+    delete_all_watches(client)

changedetectionio/tests/test_history_consistency.py

@@ -59,11 +59,29 @@ def test_consistent_history(client, live_server, measure_memory_usage, datastore
     # Wait for the sync DB save to happen
     time.sleep(2)
 
-    json_db_file = os.path.join(live_server.app.config['DATASTORE'].datastore_path, 'url-watches.json')
+    # Check which format is being used
+    datastore_path = live_server.app.config['DATASTORE'].datastore_path
+    changedetection_json = os.path.join(datastore_path, 'changedetection.json')
+    url_watches_json = os.path.join(datastore_path, 'url-watches.json')
 
-    json_obj = None
-    with open(json_db_file, 'r', encoding='utf-8') as f:
-        json_obj = json.load(f)
+    json_obj = {'watching': {}}
+
+    if os.path.exists(changedetection_json):
+        # New format: individual watch.json files
+        logger.info("Testing with new format (changedetection.json + individual watch.json)")
+
+        # Load each watch.json file
+        for uuid in live_server.app.config['DATASTORE'].data['watching'].keys():
+            watch_json_file = os.path.join(datastore_path, uuid, 'watch.json')
+            assert os.path.isfile(watch_json_file), f"watch.json should exist at {watch_json_file}"
+
+            with open(watch_json_file, 'r', encoding='utf-8') as f:
+                json_obj['watching'][uuid] = json.load(f)
+    else:
+        # Legacy format: url-watches.json
+        logger.info("Testing with legacy format (url-watches.json)")
+        with open(url_watches_json, 'r', encoding='utf-8') as f:
+            json_obj = json.load(f)
 
     # assert the right amount of watches was found in the JSON
     assert len(json_obj['watching']) == len(workers), "Correct number of watches was found in the JSON"
@@ -88,7 +106,7 @@ def test_consistent_history(client, live_server, measure_memory_usage, datastore
 
         # Find the snapshot one
         for fname in files_in_watch_dir:
-            if fname != 'history.txt' and 'html' not in fname:
+            if fname != 'history.txt' and fname != 'watch.json' and 'html' not in fname:
                 if strtobool(os.getenv("TEST_WITH_BROTLI")):
                     assert fname.endswith('.br'), "Forced TEST_WITH_BROTLI then it should be a .br filename"
 
@@ -105,13 +123,27 @@ def test_consistent_history(client, live_server, measure_memory_usage, datastore
                 assert json_obj['watching'][w]['title'], "Watch should have a title set"
                 assert contents.startswith(watch_title + "x"), f"Snapshot contents in file {fname} should start with '{watch_title}x', got '{contents}'"
 
-        assert len(files_in_watch_dir) == 3, "Should be just three files in the dir, html.br snapshot, history.txt and the extracted text snapshot"
+        # With new format, we also have watch.json, so 4 files total
+        if os.path.exists(changedetection_json):
+            assert len(files_in_watch_dir) == 4, "Should be four files in the dir with new format: watch.json, html.br snapshot, history.txt and the extracted text snapshot"
+        else:
+            assert len(files_in_watch_dir) == 3, "Should be just three files in the dir with legacy format: html.br snapshot, history.txt and the extracted text snapshot"
 
-    json_db_file = os.path.join(live_server.app.config['DATASTORE'].datastore_path, 'url-watches.json')
-    with open(json_db_file, 'r', encoding='utf-8') as f:
-        assert '"default"' not in f.read(), "'default' probably shouldnt be here, it came from when the 'default' Watch vars were accidently being saved"
+    # Check that 'default' Watch vars aren't accidentally being saved
+    if os.path.exists(changedetection_json):
+        # New format: check all individual watch.json files
+        for uuid in json_obj['watching'].keys():
+            watch_json_file = os.path.join(datastore_path, uuid, 'watch.json')
+            with open(watch_json_file, 'r', encoding='utf-8') as f:
+                assert '"default"' not in f.read(), f"'default' probably shouldnt be here in {watch_json_file}, it came from when the 'default' Watch vars were accidently being saved"
+    else:
+        # Legacy format: check url-watches.json
+        with open(url_watches_json, 'r', encoding='utf-8') as f:
+            assert '"default"' not in f.read(), "'default' probably shouldnt be here, it came from when the 'default' Watch vars were accidently being saved"
 
 
+    delete_all_watches(client)
+
 def test_check_text_history_view(client, live_server, measure_memory_usage, datastore_path):
 
     with open(os.path.join(datastore_path, "endpoint-content.txt"), "w") as f:
@@ -132,7 +164,7 @@ def test_check_text_history_view(client, live_server, measure_memory_usage, data
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
     wait_for_all_checks(client)
 
-    res = client.get(url_for("ui.ui_diff.diff_history_page", uuid="first"))
+    res = client.get(url_for("ui.ui_diff.diff_history_page", uuid=uuid))
     assert b'test-one' in res.data
     assert b'test-two' in res.data
 
@@ -150,3 +182,86 @@ def test_check_text_history_view(client, live_server, measure_memory_usage, data
     assert b'test-one' not in res.data
 
     delete_all_watches(client)
+
+
+def test_history_trim_global_only(client, live_server, measure_memory_usage, datastore_path):
+    # Add our URL to the import page
+    test_url = url_for('test_endpoint', _external=True)
+    uuid = None
+    limit = 3
+
+    for i in range(0, 10):
+        with open(os.path.join(datastore_path, "endpoint-content.txt"), "w") as f:
+            f.write(f"<html>test {i}</html>")
+        if not uuid:
+            uuid = client.application.config.get('DATASTORE').add_watch(url=test_url)
+        client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
+        wait_for_all_checks(client)
+
+        if i ==8:
+            watch = live_server.app.config['DATASTORE'].data['watching'][uuid]
+            history_n = len(list(watch.history.keys()))
+            logger.debug(f"History length should be at limit {limit} and it is {history_n}")
+            assert history_n == limit
+
+        if i == 6:
+            res = client.post(
+                url_for("settings.settings_page"),
+                data={"application-history_snapshot_max_length": limit},
+                follow_redirects=True
+            )
+            # It will need to detect one more change to start trimming it, which is really at 'start of 7'
+            assert b'Settings updated' in res.data
+
+    delete_all_watches(client)
+
+
+def test_history_trim_global_override_in_watch(client, live_server, measure_memory_usage, datastore_path):
+    # Add our URL to the import page
+    test_url = url_for('test_endpoint', _external=True)
+    uuid = None
+    limit = 3
+    res = client.post(
+        url_for("settings.settings_page"),
+        data={"application-history_snapshot_max_length": 10000},
+        follow_redirects=True
+    )
+    # It will need to detect one more change to start trimming it, which is really at 'start of 7'
+    assert b'Settings updated' in res.data
+
+
+    for i in range(0, 10):
+        with open(os.path.join(datastore_path, "endpoint-content.txt"), "w") as f:
+            f.write(f"<html>test {i}</html>")
+        if not uuid:
+            uuid = client.application.config.get('DATASTORE').add_watch(url=test_url)
+            res = client.post(
+                url_for("ui.ui_edit.edit_page", uuid="first"),
+                data={"include_filters": "", "url": test_url, "tags": "", "headers": "", 'fetch_backend': "html_requests",
+                      "time_between_check_use_default": "y", "history_snapshot_max_length": str(limit)},
+                follow_redirects=True
+            )
+            assert b"Updated watch." in res.data
+
+            wait_for_all_checks(client)
+
+        client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
+        wait_for_all_checks(client)
+
+        if i == 8:
+            watch = live_server.app.config['DATASTORE'].data['watching'][uuid]
+            history_n = len(list(watch.history.keys()))
+            logger.debug(f"History length should be at limit {limit} and it is {history_n}")
+            assert history_n == limit
+
+        if i == 6:
+            res = client.post(
+                url_for("settings.settings_page"),
+                data={"application-history_snapshot_max_length": limit},
+                follow_redirects=True
+            )
+            # It will need to detect one more change to start trimming it, which is really at 'start of 7'
+            assert b'Settings updated' in res.data
+
+    delete_all_watches(client)
+

changedetectionio/tests/test_i18n.py

@@ -160,7 +160,7 @@ def test_invalid_locale(client, live_server, measure_memory_usage, datastore_pat
 def test_language_persistence_in_session(client, live_server, measure_memory_usage, datastore_path):
     """
     Test that the language preference persists across multiple requests
-    within the same session.
+    within the same session, and that auto-detect properly clears the preference.
     """
 
     # Establish session cookie
@@ -184,6 +184,34 @@ def test_language_persistence_in_session(client, live_server, measure_memory_usa
         assert res.status_code == 200
         assert b"Annulla" in res.data, "Italian text should persist across requests"
 
+    # Verify locale is in session
+    with client.session_transaction() as sess:
+        assert sess.get('locale') == 'it', "Locale should be set in session"
+
+    # Call auto-detect to clear the locale
+    res = client.get(
+        url_for("ui.delete_locale_language_session_var_if_it_exists"),
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+    # Verify the flash message appears (in English since we cleared the locale)
+    assert b"Language set to auto-detect from browser" in res.data, "Should show flash message"
+
+    # Verify locale was removed from session
+    with client.session_transaction() as sess:
+        assert 'locale' not in sess, "Locale should be removed from session after auto-detect"
+
+    # Now requests should use browser default (English in test environment)
+    res = client.get(
+        url_for("watchlist.index"),
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+    assert b"Cancel" in res.data, "Should show English after auto-detect clears Italian"
+    assert b"Annulla" not in res.data, "Should not show Italian after auto-detect"
+
 
 def test_set_language_with_redirect(client, live_server, measure_memory_usage, datastore_path):
     """

changedetectionio/tests/test_ignorestatuscode.py

@@ -40,10 +40,7 @@ def set_some_changed_response(datastore_path):
 
 
 def test_normal_page_check_works_with_ignore_status_code(client, live_server, measure_memory_usage, datastore_path):
-
-
-    # Give the endpoint time to spin up
-    time.sleep(1)
+    from loguru import logger
 
     set_original_response(datastore_path=datastore_path)
 
@@ -62,20 +59,41 @@ def test_normal_page_check_works_with_ignore_status_code(client, live_server, me
     # Add our URL to the import page
     test_url = url_for('test_endpoint', _external=True)
     uuid = client.application.config.get('DATASTORE').add_watch(url=test_url)
+
+    logger.info(f"TEST: First check - queuing UUID {uuid}")
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
 
-    wait_for_all_checks(client)
+    logger.info(f"TEST: Waiting for first check to complete")
+    wait_result = wait_for_all_checks(client)
+    logger.info(f"TEST: First check wait completed: {wait_result}")
+
+    # Check history after first check
+    watch = client.application.config.get('DATASTORE').data['watching'][uuid]
+    logger.info(f"TEST: After first check - history count: {len(watch.history.keys())}")
 
     set_some_changed_response(datastore_path=datastore_path)
-    wait_for_all_checks(client)
+
     # Trigger a check
+    logger.info(f"TEST: Second check - queuing UUID {uuid}")
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
 
-    # Give the thread time to pick it up
-    wait_for_all_checks(client)
+    logger.info(f"TEST: Waiting for second check to complete")
+    wait_result = wait_for_all_checks(client)
+    logger.info(f"TEST: Second check wait completed: {wait_result}")
+
+    # Check history after second check
+    watch = client.application.config.get('DATASTORE').data['watching'][uuid]
+    logger.info(f"TEST: After second check - history count: {len(watch.history.keys())}")
+    logger.info(f"TEST: Watch history keys: {list(watch.history.keys())}")
 
     # It should report nothing found (no new 'has-unread-changes' class)
     res = client.get(url_for("watchlist.index"))
+
+    if b'has-unread-changes' not in res.data:
+        logger.error(f"TEST FAILED: has-unread-changes not found in response")
+        logger.error(f"TEST: Watch last_error: {watch.get('last_error')}")
+        logger.error(f"TEST: Watch last_checked: {watch.get('last_checked')}")
+
     assert b'has-unread-changes' in res.data
     assert b'/test-endpoint' in res.data
 

changedetectionio/tests/test_import.py

@@ -82,7 +82,7 @@ def test_import_distillio(client, live_server, measure_memory_usage, datastore_p
 
     # Give the endpoint time to spin up
     time.sleep(1)
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
     res = client.post(
         url_for("imports.import_page"),
         data={

changedetectionio/tests/test_jsonpath_jq_selector.py

@@ -224,6 +224,7 @@ def check_json_filter(json_filter, client, live_server, datastore_path):
     set_original_response(datastore_path=datastore_path)
 
 
+    delete_all_watches(client)
     # Add our URL to the import page
     test_url = url_for('test_endpoint', content_type="application/json", _external=True)
     uuid = client.application.config.get('DATASTORE').add_watch(url=test_url, extras={"include_filters": json_filter.splitlines()})
@@ -297,14 +298,17 @@ def check_json_filter_bool_val(json_filter, client, live_server, datastore_path)
 
 def test_check_jsonpath_filter_bool_val(client, live_server, measure_memory_usage, datastore_path):
     check_json_filter_bool_val("json:$['available']", client, live_server, datastore_path=datastore_path)
+    delete_all_watches(client)
 
 def test_check_jq_filter_bool_val(client, live_server, measure_memory_usage, datastore_path):
     if jq_support:
         check_json_filter_bool_val("jq:.available", client, live_server, datastore_path=datastore_path)
+    delete_all_watches(client)
 
 def test_check_jqraw_filter_bool_val(client, live_server, measure_memory_usage, datastore_path):
     if jq_support:
         check_json_filter_bool_val("jq:.available", client, live_server, datastore_path=datastore_path)
+    delete_all_watches(client)
 
 # Re #265 - Extended JSON selector test
 # Stuff to consider here
@@ -452,14 +456,17 @@ def test_correct_header_detect(client, live_server, measure_memory_usage, datast
 
 def test_check_jsonpath_ext_filter(client, live_server, measure_memory_usage, datastore_path):
     check_json_ext_filter('json:$[?(@.status==Sold)]', client, live_server, datastore_path=datastore_path)
+    delete_all_watches(client)
 
 def test_check_jq_ext_filter(client, live_server, measure_memory_usage, datastore_path):
     if jq_support:
         check_json_ext_filter('jq:.[] | select(.status | contains("Sold"))', client, live_server, datastore_path=datastore_path)
+    delete_all_watches(client)
 
 def test_check_jqraw_ext_filter(client, live_server, measure_memory_usage, datastore_path):
     if jq_support:
         check_json_ext_filter('jq:.[] | select(.status | contains("Sold"))', client, live_server, datastore_path=datastore_path)
+    delete_all_watches(client)
 
 def test_jsonpath_BOM_utf8(client, live_server, measure_memory_usage, datastore_path):
     from .. import html_tools
@@ -470,5 +477,6 @@ def test_jsonpath_BOM_utf8(client, live_server, measure_memory_usage, datastore_
     # See that we can find the second <script> one, which is not broken, and matches our filter
     text = html_tools.extract_json_as_string(json_str, "json:$.name")
     assert text == '"José"'
+    delete_all_watches(client)
 
     

changedetectionio/tests/test_notification.py

@@ -313,14 +313,8 @@ def test_notification_custom_endpoint_and_jinja2(client, live_server, measure_me
 
     # Add a watch and trigger a HTTP POST
     test_url = url_for('test_endpoint', _external=True)
-    res = client.post(
-        url_for("ui.ui_views.form_quick_watch_add"),
-        data={"url": test_url, "tags": 'nice one'},
-        follow_redirects=True
-    )
-
-    assert b"Watch added" in res.data
-    watch_uuid = next(iter(live_server.app.config['DATASTORE'].data['watching']))
+    watch_uuid = client.application.config.get('DATASTORE').add_watch(url=test_url, tag="nice one")
+    res = client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
 
     wait_for_all_checks(client)
     set_modified_response(datastore_path=datastore_path)

changedetectionio/tests/test_notification_errors.py

@@ -1,7 +1,7 @@
 import os
 import time
 from flask import url_for
-from .util import set_original_response, set_modified_response, live_server_setup, wait_for_all_checks
+from .util import set_original_response, set_modified_response, live_server_setup, wait_for_all_checks, delete_all_watches
 import logging
 
 def test_check_notification_error_handling(client, live_server, measure_memory_usage, datastore_path):
@@ -81,4 +81,4 @@ def test_check_notification_error_handling(client, live_server, measure_memory_u
     os.unlink(os.path.join(datastore_path, "notification.txt"))
     assert 'xxxxx' in notification_submission
 
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)

changedetectionio/tests/test_queue_handler.py

@@ -0,0 +1,52 @@
+import os
+import time
+from flask import url_for
+from .util import set_original_response,  wait_for_all_checks, wait_for_notification_endpoint_output
+from ..notification import valid_notification_formats
+from loguru import  logger
+
+def test_queue_system(client, live_server, measure_memory_usage, datastore_path):
+    """Test that multiple workers can process queue concurrently without blocking each other"""
+    # (pytest) Werkzeug's threaded server uses ThreadPoolExecutor with a default limit of around 40 threads (or min(32, os.cpu_count() + 4)).
+    items = os.cpu_count() +3
+    delay = 10
+    # Auto-queue is off here.
+    live_server.app.config['DATASTORE'].data['settings']['application']['all_paused'] = True
+
+    test_urls = [
+        f"{url_for('test_endpoint', _external=True)}?delay={delay}&id={i}&content=hello+test+content+{i}"
+        for i in range(0, items)
+    ]
+
+    # Import 30 URLs to queue
+    res = client.post(
+        url_for("imports.import_page"),
+        data={"urls": "\r\n".join(test_urls)},
+        follow_redirects=True
+    )
+    assert f"{items} Imported".encode('utf-8') in res.data
+
+    client.application.set_workers(items)
+
+    start = time.time()
+    res = client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
+    time.sleep(delay/2)
+
+    # Verify all workers are idle (no UUIDs being processed)
+    from changedetectionio import worker_pool
+    running_uuids = worker_pool.get_running_uuids()
+    logger.debug( f"Should be atleast some workers running - {len(running_uuids)} UUIDs still being processed: {running_uuids}")
+    assert len(running_uuids) != 0, f"Should be atleast some workers running - {len(running_uuids)} UUIDs still being processed: {running_uuids}"
+
+    wait_for_all_checks(client)
+
+    # all workers should be done in less than say 10 seconds (they take time to 'see' something is in the queue too)
+    total_time = (time.time() - start)
+    logger.debug(f"All workers finished {items} items in less than {delay} seconds per job. {total_time}s total")
+    # if there was a bug in queue handler not running parallel, this would blow out to items*delay seconds
+    assert total_time < delay + 10, f"All workers finished {items} items in less than {delay} seconds per job, total time {total_time}s"
+
+    # Verify all workers are idle (no UUIDs being processed)
+    from changedetectionio import worker_pool
+    running_uuids = worker_pool.get_running_uuids()
+    assert len(running_uuids) == 0, f"Expected all workers to be idle, but {len(running_uuids)} UUIDs still being processed: {running_uuids}"

changedetectionio/tests/test_request.py

@@ -17,12 +17,12 @@ def test_headers_in_request(client, live_server, measure_memory_usage, datastore
         test_url = test_url.replace('localhost', 'changedet')
 
     # Add the test URL twice, we will check
-    uuid = client.application.config.get('DATASTORE').add_watch(url=test_url)
+    uuidA = client.application.config.get('DATASTORE').add_watch(url=test_url)
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
 
     wait_for_all_checks(client)
 
-    uuid = client.application.config.get('DATASTORE').add_watch(url=test_url)
+    uuidB = client.application.config.get('DATASTORE').add_watch(url=test_url)
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
 
     wait_for_all_checks(client)
@@ -31,7 +31,7 @@ def test_headers_in_request(client, live_server, measure_memory_usage, datastore
 
     # Add some headers to a request
     res = client.post(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuidA),
         data={
               "url": test_url,
               "tags": "",
@@ -42,13 +42,14 @@ def test_headers_in_request(client, live_server, measure_memory_usage, datastore
     )
     assert b"Updated watch." in res.data
 
+    client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
 
     # Give the thread time to pick up the first version
     wait_for_all_checks(client)
 
     # The service should echo back the request headers
     res = client.get(
-        url_for("ui.ui_preview.preview_page", uuid="first"),
+        url_for("ui.ui_preview.preview_page", uuid=uuidA),
         follow_redirects=True
     )
 
@@ -92,7 +93,7 @@ def test_body_in_request(client, live_server, measure_memory_usage, datastore_pa
 
     # add the first 'version'
     res = client.post(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuid),
         data={
               "url": test_url,
               "tags": "",
@@ -110,7 +111,7 @@ def test_body_in_request(client, live_server, measure_memory_usage, datastore_pa
     body_value = 'Test Body Value {{ 1+1 }}'
     body_value_formatted = 'Test Body Value 2'
     res = client.post(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuid),
         data={
               "url": test_url,
               "tags": "",
@@ -126,7 +127,7 @@ def test_body_in_request(client, live_server, measure_memory_usage, datastore_pa
 
     # The service should echo back the body
     res = client.get(
-        url_for("ui.ui_preview.preview_page", uuid="first"),
+        url_for("ui.ui_preview.preview_page", uuid=uuid),
         follow_redirects=True
     )
 
@@ -142,10 +143,14 @@ def test_body_in_request(client, live_server, measure_memory_usage, datastore_pa
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
     wait_for_all_checks(client)
     watches_with_body = 0
-    with open(os.path.join(datastore_path, 'url-watches.json'), encoding='utf-8') as f:
-        app_struct = json.load(f)
-        for uuid in app_struct['watching']:
-            if app_struct['watching'][uuid]['body']==body_value:
+
+    # Read individual watch.json files
+    for uuid in client.application.config.get('DATASTORE').data['watching'].keys():
+        watch_json_file = os.path.join(datastore_path, uuid, 'watch.json')
+        assert os.path.exists(watch_json_file), f"watch.json should exist at {watch_json_file}"
+        with open(watch_json_file, 'r', encoding='utf-8') as f:
+            watch_data = json.load(f)
+            if watch_data.get('body') == body_value:
                 watches_with_body += 1
 
     # Should be only one with body set
@@ -153,7 +158,7 @@ def test_body_in_request(client, live_server, measure_memory_usage, datastore_pa
 
     # Attempt to add a body with a GET method
     res = client.post(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuid),
         data={
               "url": test_url,
               "tags": "",
@@ -225,10 +230,14 @@ def test_method_in_request(client, live_server, measure_memory_usage, datastore_
     wait_for_all_checks(client)
 
     watches_with_method = 0
-    with open(os.path.join(datastore_path, 'url-watches.json'), encoding='utf-8') as f:
-        app_struct = json.load(f)
-        for uuid in app_struct['watching']:
-            if app_struct['watching'][uuid]['method'] == 'PATCH':
+
+    # Read individual watch.json files
+    for uuid in client.application.config.get('DATASTORE').data['watching'].keys():
+        watch_json_file = os.path.join(datastore_path, uuid, 'watch.json')
+        assert os.path.exists(watch_json_file), f"watch.json should exist at {watch_json_file}"
+        with open(watch_json_file, 'r', encoding='utf-8') as f:
+            watch_data = json.load(f)
+            if watch_data.get('method') == 'PATCH':
                 watches_with_method += 1
 
     # Should be only one with method set to PATCH

changedetectionio/tests/test_restock_itemprop.py

@@ -236,6 +236,7 @@ def test_restock_itemprop_with_tag(client, live_server, measure_memory_usage, da
     }
 
     _run_test_minmax_limit(client, extra_watch_edit_form=extras,datastore_path=datastore_path)
+    delete_all_watches(client)
 
 
 
@@ -388,9 +389,10 @@ def test_change_with_notification_values(client, live_server, measure_memory_usa
     os.unlink(os.path.join(datastore_path, "notification.txt"))
     uuid = next(iter(live_server.app.config['DATASTORE'].data['watching']))
     res = client.post(url_for("ui.ui_notification.ajax_callback_send_notification_test", watch_uuid=uuid), data={}, follow_redirects=True)
-    time.sleep(5)
+    wait_for_notification_endpoint_output(datastore_path=datastore_path)
     assert os.path.isfile(os.path.join(datastore_path, "notification.txt")), "Notification received"
 
+    delete_all_watches(client)
 
 def test_data_sanity(client, live_server, measure_memory_usage, datastore_path):
     
@@ -406,6 +408,7 @@ def test_data_sanity(client, live_server, measure_memory_usage, datastore_path):
         follow_redirects=True
     )
 
+    client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
 
     wait_for_all_checks(client)
     res = client.get(url_for("watchlist.index"))
@@ -417,6 +420,7 @@ def test_data_sanity(client, live_server, measure_memory_usage, datastore_path):
         data={"url": test_url2, "tags": 'restock tests', 'processor': 'restock_diff'},
         follow_redirects=True
     )
+    client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
     wait_for_all_checks(client)
     res = client.get(url_for("watchlist.index"))
     assert str(res.data.decode()).count("950.95") == 1, "Price should only show once (for the watch added, no other watches yet)"
@@ -462,3 +466,4 @@ def test_special_prop_examples(client, live_server, measure_memory_usage, datast
             assert b'ception' not in res.data
             assert b'155.55' in res.data
 
+    delete_all_watches(client)

changedetectionio/tests/test_rss.py

@@ -107,7 +107,7 @@ def test_rss_and_token(client, live_server, measure_memory_usage, datastore_path
     assert b"Access denied, bad token" not in res.data
     assert b"Random content" in res.data
 
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 def test_basic_cdata_rss_markup(client, live_server, measure_memory_usage, datastore_path):
     

changedetectionio/tests/test_rss_single_watch.py

@@ -23,6 +23,7 @@ def test_rss_feed_empty(client, live_server, measure_memory_usage, datastore_pat
     )
     assert res.status_code == 400
     assert b'does not have enough history snapshots to show' in res.data
+    delete_all_watches(client)
 
 def test_rss_single_watch_order(client, live_server, measure_memory_usage, datastore_path):
     """

changedetectionio/tests/test_share_watch.py

@@ -24,20 +24,20 @@ def test_share_watch(client, live_server, measure_memory_usage, datastore_path):
     # Goto the edit page, add our ignore text
     # Add our URL to the import page
     res = client.post(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuid),
         data={"include_filters": include_filters, "url": test_url, "tags": "", "headers": "", 'fetch_backend': "html_requests", "time_between_check_use_default": "y"},
         follow_redirects=True
     )
     assert b"Updated watch." in res.data
     # Check it saved
     res = client.get(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuid),
     )
     assert bytes(include_filters.encode('utf-8')) in res.data
 
     # click share the link
     res = client.get(
-        url_for("ui.form_share_put_watch", uuid="first"),
+        url_for("ui.form_share_put_watch", uuid=uuid),
         follow_redirects=True
     )
 
@@ -63,13 +63,16 @@ def test_share_watch(client, live_server, measure_memory_usage, datastore_path):
 
     # Now hit edit, we should see what we expect
     # that the import fetched the meta-data
-
+    uuids = list(client.application.config.get('DATASTORE').data['watching'])
+    assert uuids, "It saved/imported and created a new URL from the share"
     # Check it saved
     res = client.get(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuids[0]),
     )
     assert bytes(include_filters.encode('utf-8')) in res.data
 
     # Check it saved the URL
     res = client.get(url_for("watchlist.index"))
     assert bytes(test_url.encode('utf-8')) in res.data
+
+    delete_all_watches(client)

changedetectionio/tests/test_ui.py

@@ -25,6 +25,7 @@ def test_recheck_time_field_validation_global_settings(client, live_server, meas
 
 
     assert REQUIRE_ATLEAST_ONE_TIME_PART_MESSAGE_DEFAULT.encode('utf-8') in res.data
+    delete_all_watches(client)
 
 
 def test_recheck_time_field_validation_single_watch(client, live_server, measure_memory_usage, datastore_path):
@@ -94,6 +95,7 @@ def test_recheck_time_field_validation_single_watch(client, live_server, measure
 
     assert b"Updated watch." in res.data
     assert REQUIRE_ATLEAST_ONE_TIME_PART_WHEN_NOT_GLOBAL_DEFAULT.encode('utf-8') not in res.data
+    delete_all_watches(client)
 
 def test_checkbox_open_diff_in_new_tab(client, live_server, measure_memory_usage, datastore_path):
     
@@ -242,6 +244,7 @@ def test_page_title_listing_behaviour(client, live_server, measure_memory_usage,
     # No page title description, and 'use_page_title_in_list' is on, it should show the <title>
     res = client.get(url_for("watchlist.index"))
     assert b"head titlecustom html" in res.data
+    delete_all_watches(client)
 
 
 def test_ui_viewed_unread_flag(client, live_server, measure_memory_usage, datastore_path):
@@ -283,4 +286,5 @@ def test_ui_viewed_unread_flag(client, live_server, measure_memory_usage, datast
     client.get(url_for("ui.mark_all_viewed"), follow_redirects=True)
     time.sleep(0.2)
     res = client.get(url_for("watchlist.index"))
-    assert b'<span id="unread-tab-counter">0</span>' in res.data
+    assert b'<span id="unread-tab-counter">0</span>' in res.data
+    delete_all_watches(client)

changedetectionio/tests/test_xpath_selector.py

@@ -366,7 +366,7 @@ def test_check_with_prefix_include_filters(client, live_server, measure_memory_u
     assert b"Some text thats the same" in res.data  # in selector
     assert b"Some text that will change" not in res.data  # not in selector
 
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 
 def test_various_rules(client, live_server, measure_memory_usage, datastore_path):
@@ -423,7 +423,7 @@ def test_xpath_20(client, live_server, measure_memory_usage, datastore_path):
 
     test_url = url_for('test_endpoint', _external=True)
     res = client.post(
-        url_for("ui.ui_edit.edit_page", uuid="first"),
+        url_for("ui.ui_edit.edit_page", uuid=uuid),
         data={"include_filters": "//*[contains(@class, 'sametext')]|//*[contains(@class, 'changetext')]",
               "url": test_url,
               "tags": "",
@@ -437,14 +437,14 @@ def test_xpath_20(client, live_server, measure_memory_usage, datastore_path):
     wait_for_all_checks(client)
 
     res = client.get(
-        url_for("ui.ui_preview.preview_page", uuid="first"),
+        url_for("ui.ui_preview.preview_page", uuid=uuid),
         follow_redirects=True
     )
 
     assert b"Some text thats the same" in res.data  # in selector
     assert b"Some text that will change" in res.data  # in selector
 
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 
 def test_xpath_20_function_count(client, live_server, measure_memory_usage, datastore_path):
@@ -477,7 +477,7 @@ def test_xpath_20_function_count(client, live_server, measure_memory_usage, data
 
     assert b"246913579975308642" in res.data  # in selector
 
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 
 def test_xpath_20_function_count2(client, live_server, measure_memory_usage, datastore_path):
@@ -501,6 +501,8 @@ def test_xpath_20_function_count2(client, live_server, measure_memory_usage, dat
     )
 
     assert b"Updated watch." in res.data
+    client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
+
     wait_for_all_checks(client)
 
     res = client.get(
@@ -510,7 +512,7 @@ def test_xpath_20_function_count2(client, live_server, measure_memory_usage, dat
 
     assert b"246913579975308642" in res.data  # in selector
 
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 
 def test_xpath_20_function_string_join_matches(client, live_server, measure_memory_usage, datastore_path):
@@ -544,7 +546,7 @@ def test_xpath_20_function_string_join_matches(client, live_server, measure_memo
 
     assert b"Some text thats the samespecialconjunctionSome text that will change" in res.data  # in selector
 
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 
 def _subtest_xpath_rss(client, datastore_path, content_type='text/html'):
@@ -582,7 +584,7 @@ def _subtest_xpath_rss(client, datastore_path, content_type='text/html'):
     assert b"Lets go discount" in res.data, f"When testing for Lets go discount called with content type '{content_type}'"
     assert b"Events and Announcements" not in res.data, f"When testing for Lets go discount called with content type '{content_type}'" # It should not be here because thats not our selector target
 
-    client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
+    delete_all_watches(client)
 
 # Be sure all-in-the-wild types of RSS feeds work with xpath
 def test_rss_xpath(client, live_server, measure_memory_usage, datastore_path):

changedetectionio/tests/unit/test_watch_model.py

@@ -5,15 +5,24 @@
 
 import unittest
 import os
+import pickle
+from copy import deepcopy
 
-from changedetectionio.model import Watch
+from changedetectionio.model import Watch, Tag
 
 # mostly
 class TestDiffBuilder(unittest.TestCase):
 
     def test_watch_get_suggested_from_diff_timestamp(self):
         import uuid as uuid_builder
-        watch = Watch.model(datastore_path='/tmp', default={})
+        # Create minimal mock datastore for tests
+        mock_datastore = {
+            'settings': {
+                'application': {}
+            },
+            'watching': {}
+        }
+        watch = Watch.model(datastore_path='/tmp', __datastore=mock_datastore, default={})
         watch.ensure_data_dir_exists()
 
 
@@ -49,7 +58,7 @@ class TestDiffBuilder(unittest.TestCase):
         assert p == "109", "Correct when its the same time"
 
         # new empty one
-        watch = Watch.model(datastore_path='/tmp', default={})
+        watch = Watch.model(datastore_path='/tmp', __datastore=mock_datastore, default={})
         p = watch.get_from_version_based_on_last_viewed
         assert p == None, "None when no history available"
 
@@ -61,5 +70,184 @@ class TestDiffBuilder(unittest.TestCase):
         p = watch.get_from_version_based_on_last_viewed
         assert p == "100", "Correct with only one history snapshot"
 
+    def test_watch_deepcopy_doesnt_copy_datastore(self):
+        """
+        CRITICAL: Ensure deepcopy(watch) shares __datastore instead of copying it.
+
+        Without this, deepcopy causes exponential memory growth:
+        - 100 watches × deepcopy each = 10,000 watch objects in memory (100²)
+        - Memory grows from 120MB → 2GB
+
+        This test prevents regressions in the __deepcopy__ implementation.
+        """
+        # Create mock datastore with multiple watches
+        mock_datastore = {
+            'settings': {'application': {'history_snapshot_max_length': 10}},
+            'watching': {}
+        }
+
+        # Create 3 watches that all reference the same datastore
+        watches = []
+        for i in range(3):
+            watch = Watch.model(
+                __datastore=mock_datastore,
+                datastore_path='/tmp/test',
+                default={'url': f'https://example{i}.com', 'title': f'Watch {i}'}
+            )
+            mock_datastore['watching'][watch['uuid']] = watch
+            watches.append(watch)
+
+        # Test 1: Deepcopy shares datastore reference (doesn't copy it)
+        watch_copy = deepcopy(watches[0])
+
+        self.assertIsNotNone(watch_copy._datastore,
+                            "__datastore should exist in copied watch")
+        self.assertIs(watch_copy._datastore, watches[0]._datastore,
+                     "__datastore should be SHARED (same object), not copied")
+        self.assertIs(watch_copy._datastore, mock_datastore,
+                     "__datastore should reference the original datastore")
+
+        # Test 2: Dict data is properly copied (not shared)
+        self.assertEqual(watch_copy['title'], 'Watch 0', "Dict data should be copied")
+        watch_copy['title'] = 'MODIFIED'
+        self.assertNotEqual(watches[0]['title'], 'MODIFIED',
+                           "Modifying copy should not affect original")
+
+        # Test 3: Verify no nested datastore copies in watch dict
+        # The dict should only contain watch settings, not the datastore
+        watch_dict = dict(watch_copy)
+        self.assertNotIn('__datastore', watch_dict,
+                        "__datastore should not be in dict keys")
+        self.assertNotIn('_model__datastore', watch_dict,
+                        "_model__datastore should not be in dict keys")
+
+        # Test 4: Multiple deepcopies don't cause exponential memory growth
+        # If datastore was copied, each copy would contain 3 watches,
+        # and those watches would contain the datastore, etc. (infinite recursion)
+        copies = []
+        for _ in range(5):
+            copies.append(deepcopy(watches[0]))
+
+        # All copies should share the same datastore
+        for copy in copies:
+            self.assertIs(copy._datastore, mock_datastore,
+                         "All copies should share the original datastore")
+
+    def test_watch_pickle_doesnt_serialize_datastore(self):
+        """
+        Ensure pickle/unpickle doesn't serialize __datastore.
+
+        This is important for multiprocessing and caching - we don't want
+        to serialize the entire datastore when pickling a watch.
+        """
+        mock_datastore = {
+            'settings': {'application': {}},
+            'watching': {}
+        }
+
+        watch = Watch.model(
+            __datastore=mock_datastore,
+            datastore_path='/tmp/test',
+            default={'url': 'https://example.com', 'title': 'Test Watch'}
+        )
+
+        # Pickle and unpickle
+        pickled = pickle.dumps(watch)
+        unpickled_watch = pickle.loads(pickled)
+
+        # Test 1: Watch data is preserved
+        self.assertEqual(unpickled_watch['url'], 'https://example.com',
+                        "Dict data should be preserved after pickle/unpickle")
+
+        # Test 2: __datastore is NOT serialized (attribute shouldn't exist after unpickle)
+        self.assertFalse(hasattr(unpickled_watch, '_datastore'),
+                         "__datastore attribute should not exist after unpickle (not serialized)")
+
+        # Test 3: Pickled data shouldn't contain the large datastore object
+        # If datastore was serialized, the pickle size would be much larger
+        pickle_size = len(pickled)
+        # A single watch should be small (< 10KB), not include entire datastore
+        self.assertLess(pickle_size, 10000,
+                       f"Pickled watch too large ({pickle_size} bytes) - might include datastore")
+
+    def test_tag_deepcopy_works(self):
+        """
+        Ensure Tag objects (which also inherit from watch_base) can be deepcopied.
+
+        Tags now have optional __datastore for consistency with Watch objects.
+        """
+        mock_datastore = {
+            'settings': {'application': {}},
+            'watching': {}
+        }
+
+        # Test 1: Tag without datastore (backward compatibility)
+        tag_without_ds = Tag.model(
+            datastore_path='/tmp/test',
+            default={'title': 'Test Tag', 'overrides_watch': True}
+        )
+        tag_copy1 = deepcopy(tag_without_ds)
+        self.assertEqual(tag_copy1['title'], 'Test Tag', "Tag data should be copied")
+
+        # Test 2: Tag with datastore (new pattern for consistency)
+        tag_with_ds = Tag.model(
+            datastore_path='/tmp/test',
+            __datastore=mock_datastore,
+            default={'title': 'Test Tag With DS', 'overrides_watch': True}
+        )
+
+        # Deepcopy should work
+        tag_copy2 = deepcopy(tag_with_ds)
+
+        # Test 3: Dict data is copied
+        self.assertEqual(tag_copy2['title'], 'Test Tag With DS', "Tag data should be copied")
+
+        # Test 4: Modifications to copy don't affect original
+        tag_copy2['title'] = 'MODIFIED'
+        self.assertNotEqual(tag_with_ds['title'], 'MODIFIED',
+                           "Modifying copy should not affect original")
+
+        # Test 5: Tag with datastore shares it (doesn't copy it)
+        if hasattr(tag_with_ds, '_datastore'):
+            self.assertIs(tag_copy2._datastore, tag_with_ds._datastore,
+                         "Tag should share __datastore reference like Watch does")
+
+    def test_watch_copy_performance(self):
+        """
+        Verify that our __deepcopy__ implementation doesn't cause performance issues.
+
+        With the fix, deepcopy should be fast because we're sharing datastore
+        instead of copying it.
+        """
+        import time
+
+        # Create a watch with large datastore (many watches)
+        mock_datastore = {
+            'settings': {'application': {}},
+            'watching': {}
+        }
+
+        # Add 100 watches to the datastore
+        for i in range(100):
+            w = Watch.model(
+                __datastore=mock_datastore,
+                datastore_path='/tmp/test',
+                default={'url': f'https://example{i}.com'}
+            )
+            mock_datastore['watching'][w['uuid']] = w
+
+        # Time how long deepcopy takes
+        watch = list(mock_datastore['watching'].values())[0]
+
+        start = time.time()
+        for _ in range(10):
+            _ = deepcopy(watch)
+        elapsed = time.time() - start
+
+        # Should be fast (< 0.1 seconds for 10 copies)
+        # If datastore was copied, it would take much longer
+        self.assertLess(elapsed, 0.5,
+                       f"Deepcopy too slow ({elapsed:.3f}s for 10 copies) - might be copying datastore")
+
 if __name__ == '__main__':
     unittest.main()

changedetectionio/tests/util.py

@@ -6,6 +6,42 @@ from flask import url_for
 import logging
 import time
 import os
+import threading
+
+# Thread-safe global storage for test endpoint content
+# Avoids filesystem cache issues in parallel tests
+_test_endpoint_content_lock = threading.Lock()
+_test_endpoint_content = {}
+
+def write_test_file_and_sync(filepath, content, mode='w'):
+    """
+    Write test data to file and ensure it's synced to disk.
+    Also stores in thread-safe global dict to bypass filesystem cache.
+
+    Critical for parallel tests where workers may read files immediately after write.
+    Without fsync(), data may still be in OS buffers when workers try to read,
+    causing race conditions where old data is seen.
+
+    Args:
+        filepath: Full path to file
+        content: Content to write (str or bytes)
+        mode: File mode ('w' for text, 'wb' for binary)
+    """
+    # Convert content to bytes if needed
+    if isinstance(content, str):
+        content_bytes = content.encode('utf-8')
+    else:
+        content_bytes = content
+
+    # Store in thread-safe global dict for instant access
+    with _test_endpoint_content_lock:
+        _test_endpoint_content[os.path.basename(filepath)] = content_bytes
+
+    # Also write to file for compatibility
+    with open(filepath, mode) as f:
+        f.write(content)
+        f.flush()  # Flush Python buffer to OS
+        os.fsync(f.fileno())  # Force OS to write to disk
 
 def set_original_response(datastore_path, extra_title=''):
     test_return_data = f"""<html>
@@ -20,8 +56,7 @@ def set_original_response(datastore_path, extra_title=''):
      </html>
     """
 
-    with open(os.path.join(datastore_path, "endpoint-content.txt"), "w") as f:
-        f.write(test_return_data)
+    write_test_file_and_sync(os.path.join(datastore_path, "endpoint-content.txt"), test_return_data)
     return None
 
 def set_modified_response(datastore_path):
@@ -36,9 +71,7 @@ def set_modified_response(datastore_path):
      </html>
     """
 
-    with open(os.path.join(datastore_path, "endpoint-content.txt"), "w") as f:
-        f.write(test_return_data)
-
+    write_test_file_and_sync(os.path.join(datastore_path, "endpoint-content.txt"), test_return_data)
     return None
 def set_longer_modified_response(datastore_path):
     test_return_data = """<html>
@@ -55,9 +88,7 @@ def set_longer_modified_response(datastore_path):
      </html>
     """
 
-    with open(os.path.join(datastore_path, "endpoint-content.txt"), "w") as f:
-        f.write(test_return_data)
-
+    write_test_file_and_sync(os.path.join(datastore_path, "endpoint-content.txt"), test_return_data)
     return None
 
 def set_more_modified_response(datastore_path):
@@ -73,17 +104,14 @@ def set_more_modified_response(datastore_path):
      </html>
     """
 
-    with open(os.path.join(datastore_path, "endpoint-content.txt"), "w") as f:
-        f.write(test_return_data)
-
+    write_test_file_and_sync(os.path.join(datastore_path, "endpoint-content.txt"), test_return_data)
     return None
 
 
 def set_empty_text_response(datastore_path):
     test_return_data = """<html><body></body></html>"""
 
-    with open(os.path.join(datastore_path, "endpoint-content.txt"), "w") as f:
-        f.write(test_return_data)
+    write_test_file_and_sync(os.path.join(datastore_path, "endpoint-content.txt"), test_return_data)
 
     return None
 
@@ -132,46 +160,35 @@ def extract_UUID_from_client(client):
     return uuid.strip()
 
 def delete_all_watches(client=None):
+
     uuids = list(client.application.config.get('DATASTORE').data['watching'])
     for uuid in uuids:
         client.application.config.get('DATASTORE').delete(uuid)
+    from changedetectionio.flask_app import update_q
 
+    # Clear the queue to prevent leakage to next test
+    # Use clear() method to ensure both priority_items and notification_queue are drained
+    if hasattr(update_q, 'clear'):
+        update_q.clear()
+    else:
+        # Fallback for old implementation
+        while not update_q.empty():
+            try:
+                update_q.get_nowait()
+            except:
+                break
+
+    time.sleep(0.2)
 
 def wait_for_all_checks(client=None):
     """
     Waits until the queue is empty and workers are idle.
-    Much faster than the original with adaptive timing.
+    Delegates to worker_pool.wait_for_all_checks for shared logic.
     """
     from changedetectionio.flask_app import update_q as global_update_q
-    from changedetectionio import worker_handler
-    empty_since = None
-    attempt = 0
-    max_attempts = 150  # Still reasonable upper bound
+    from changedetectionio import worker_pool
+    return worker_pool.wait_for_all_checks(global_update_q, timeout=150)
 
-    while attempt < max_attempts:
-        # Start with fast checks, slow down if needed
-        if attempt < 10:
-            time.sleep(0.2)  # Very fast initial checks
-        elif attempt < 30:
-            time.sleep(0.4)  # Medium speed
-        else:
-            time.sleep(0.8)  # Slower for persistent issues
-
-        q_length = global_update_q.qsize()
-        running_uuids = worker_handler.get_running_uuids()
-        any_workers_busy = len(running_uuids) > 0
-
-        if q_length == 0 and not any_workers_busy:
-            if empty_since is None:
-                empty_since = time.time()
-            # Brief stabilization period for async workers
-            elif time.time() - empty_since >= 0.3:
-                break
-        else:
-            empty_since = None
-
-        attempt += 1
-        time.sleep(0.3)
 
 def wait_for_watch_history(client, min_history_count=2, timeout=10):
     """
@@ -220,8 +237,11 @@ def new_live_server_setup(live_server):
 
     @live_server.app.route('/test-endpoint')
     def test_endpoint():
-        from loguru import logger
-        logger.debug(f"/test-endpoint hit {request}")
+        # REMOVED: logger.debug() causes file locking between test process and Flask server process
+        # Flask server runs in separate multiprocessing.Process and inherited loguru tries to
+        # write to same log files, causing request handlers to block on file locks
+        # from loguru import logger
+        # logger.debug(f"/test-endpoint hit {request}")
         ctype = request.args.get('content_type')
         status_code = request.args.get('status_code')
         content = request.args.get('content') or None
@@ -243,15 +263,35 @@ def new_live_server_setup(live_server):
                     resp.headers['Content-Type'] = ctype if ctype else 'text/html'
                 return resp
 
-            # Tried using a global var here but didn't seem to work, so reading from a file instead.
-            datastore_path = current_app.config.get('TEST_DATASTORE_PATH', 'test-datastore')
-            with open(os.path.join(datastore_path, "endpoint-content.txt"), "rb") as f:
-                resp = make_response(f.read(), status_code)
-                if uppercase_headers:
-                    resp.headers['CONTENT-TYPE'] = ctype if ctype else 'text/html'
-                else:
-                    resp.headers['Content-Type'] = ctype if ctype else 'text/html'
-                return resp
+            # Check thread-safe global dict first (instant, no cache issues)
+            # Fall back to file if not in dict (for tests that write directly)
+            with _test_endpoint_content_lock:
+                content_data = _test_endpoint_content.get("endpoint-content.txt")
+
+            if content_data is None:
+                # Not in global dict, read from file
+                datastore_path = current_app.config.get('TEST_DATASTORE_PATH', 'test-datastore')
+                filepath = os.path.join(datastore_path, "endpoint-content.txt")
+
+                # REMOVED: os.sync() was blocking for many seconds during parallel tests
+                # With -n 6+ parallel tests, heavy I/O causes os.sync() to wait for ALL
+                # system writes to complete, causing "Read timed out" errors
+                # File writes from test code are already flushed by the time workers fetch
+
+                try:
+                    with open(filepath, "rb") as f:
+                        content_data = f.read()
+                except Exception as e:
+                    # REMOVED: logger.error() causes file locking in multiprocess context
+                    # Just raise the exception directly for debugging
+                    raise
+
+            resp = make_response(content_data, status_code)
+            if uppercase_headers:
+                resp.headers['CONTENT-TYPE'] = ctype if ctype else 'text/html'
+            else:
+                resp.headers['Content-Type'] = ctype if ctype else 'text/html'
+            return resp
         except FileNotFoundError:
             return make_response('', status_code)
 
@@ -326,6 +366,12 @@ def new_live_server_setup(live_server):
     def test_pdf_endpoint():
         datastore_path = current_app.config.get('TEST_DATASTORE_PATH', 'test-datastore')
 
+        # Force filesystem sync before reading to ensure fresh data
+        try:
+            os.sync()
+        except (AttributeError, PermissionError):
+            pass
+
         # Tried using a global var here but didn't seem to work, so reading from a file instead.
         with open(os.path.join(datastore_path, "endpoint-test.pdf"), "rb") as f:
             resp = make_response(f.read(), 200)

changedetectionio/translations/README.md

@@ -0,0 +1,101 @@
+# Translation Guide
+
+## Updating Translations
+
+To maintain consistency and minimize unnecessary changes in translation files, run these commands:
+
+```bash
+python setup.py extract_messages   # Extract translatable strings
+python setup.py update_catalog     # Update all language files
+python setup.py compile_catalog    # Compile to binary .mo files
+```
+
+## Configuration
+
+All translation settings are configured in **`../../setup.cfg`** (single source of truth).
+
+The configuration below is shown for reference - **edit `setup.cfg` to change settings**:
+
+```ini
+[extract_messages]
+# Extract translatable strings from source code
+mapping_file = babel.cfg
+output_file = changedetectionio/translations/messages.pot
+input_paths = changedetectionio
+keywords = _ _l gettext
+# Options to reduce unnecessary changes in .pot files
+sort_by_file = true       # Keeps entries ordered by file path
+width = 120               # Consistent line width (prevents rewrapping)
+add_location = file       # Show file path only (not line numbers)
+
+[update_catalog]
+# Update existing .po files with new strings from .pot
+# Note: 'locale' is omitted - Babel auto-discovers all catalogs in output_dir
+input_file = changedetectionio/translations/messages.pot
+output_dir = changedetectionio/translations
+domain = messages
+# Options for consistent formatting
+width = 120               # Consistent line width
+no_fuzzy_matching = true  # Avoids incorrect automatic matches
+
+[compile_catalog]
+# Compile .po files to .mo binary format
+directory = changedetectionio/translations
+domain = messages
+```
+
+**Key formatting options:**
+- `sort_by_file = true` - Orders entries by file path (consistent ordering)
+- `width = 120` - Fixed line width prevents text rewrapping
+- `add_location = file` - Shows file path only, not line numbers (reduces git churn)
+- `no_fuzzy_matching = true` - Prevents incorrect automatic fuzzy matches
+
+## Why Use These Commands?
+
+Running pybabel commands directly without consistent options causes:
+- ❌ Entries get reordered differently each time
+- ❌ Text gets rewrapped at different widths
+- ❌ Line numbers change every edit (if not configured)
+- ❌ Large diffs that make code review difficult
+
+Using `python setup.py` commands ensures:
+- ✅ Consistent ordering (by file path, not alphabetically)
+- ✅ Consistent line width (120 characters, no rewrapping)
+- ✅ File-only locations (no line number churn)
+- ✅ No fuzzy matching (prevents incorrect auto-translations)
+- ✅ Minimal diffs (only actual changes show up)
+- ✅ Easier code review and git history
+
+These commands read settings from `../../setup.cfg` automatically.
+
+## Supported Languages
+
+- `cs` - Czech (Čeština)
+- `de` - German (Deutsch)
+- `en_GB` - English (UK)
+- `en_US` - English (US)
+- `fr` - French (Français)
+- `it` - Italian (Italiano)
+- `ko` - Korean (한국어)
+- `zh` - Chinese Simplified (中文简体)
+- `zh_Hant_TW` - Chinese Traditional (繁體中文)
+
+## Adding a New Language
+
+1. Initialize the new language catalog:
+   ```bash
+   pybabel init -i changedetectionio/translations/messages.pot -d changedetectionio/translations -l NEW_LANG_CODE
+   ```
+2. Compile it:
+   ```bash
+   python setup.py compile_catalog
+   ```
+
+Babel will auto-discover the new language on subsequent translation updates.
+
+## Translation Notes
+
+From CLAUDE.md:
+- Always use "monitor" or "watcher" terminology (not "clock")
+- Use the most brief wording suitable
+- When finding issues in one language, check ALL languages for the same issue