add note

.github/workflows/containers.yml

@@ -15,7 +15,6 @@ on:
   push:
     branches:
       - master
-      - dev
 
 jobs:
   metadata:
@@ -93,17 +92,28 @@ jobs:
           version: latest
           driver-opts: image=moby/buildkit:master
 
-      # dev branch -> :dev container tag
+      # master branch -> :dev container tag
+      - name: Docker meta :dev
+        if: ${{ github.ref == 'refs/heads/master' && github.event_name != 'release' }}
+        uses: docker/metadata-action@v5
+        id: meta_dev
+        with:
+          images: |
+            ${{ secrets.DOCKER_HUB_USERNAME }}/changedetection.io
+            ghcr.io/${{ github.repository }}
+          tags: |
+            type=raw,value=dev
+
       - name: Build and push :dev
         id: docker_build
-        if: ${{ github.ref == 'refs/heads/dev' }}
+        if: ${{ github.ref == 'refs/heads/master' && github.event_name != 'release' }}
         uses: docker/build-push-action@v6
         with:
           context: ./
           file: ./Dockerfile
           push: true
-          tags: |
-            ${{ secrets.DOCKER_HUB_USERNAME }}/changedetection.io:dev,ghcr.io/${{ github.repository }}:dev
+          tags: ${{ steps.meta_dev.outputs.tags }}
+          labels: ${{ steps.meta_dev.outputs.labels }}
           platforms: linux/amd64,linux/arm64,linux/arm/v7,linux/arm/v8
           cache-from: type=gha
           cache-to: type=gha,mode=max
@@ -142,6 +152,7 @@ jobs:
           file: ./Dockerfile
           push: true
           tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
           platforms: linux/amd64,linux/arm64,linux/arm/v7,linux/arm/v8
           cache-from: type=gha
           cache-to: type=gha,mode=max

.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 }} ...
@@ -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

.gitignore

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

Dockerfile

@@ -78,6 +78,12 @@ RUN --mount=type=cache,id=pip,sharing=locked,target=/tmp/pip-cache \
 # Final image stage
 FROM python:${PYTHON_VERSION}-slim-bookworm
 LABEL org.opencontainers.image.source="https://github.com/dgtlmoon/changedetection.io"
+LABEL org.opencontainers.image.url="https://changedetection.io"
+LABEL org.opencontainers.image.documentation="https://changedetection.io/tutorials"
+LABEL org.opencontainers.image.title="changedetection.io"
+LABEL org.opencontainers.image.description="Self-hosted web page change monitoring and notification service"
+LABEL org.opencontainers.image.licenses="Apache-2.0"
+LABEL org.opencontainers.image.vendor="changedetection.io"
 
 RUN apt-get update && apt-get install -y --no-install-recommends \
     libxslt1.1 \
@@ -132,6 +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 *
@@ -16,6 +17,7 @@ recursive-include changedetectionio/widgets *
 prune changedetectionio/static/package-lock.json
 prune changedetectionio/static/styles/node_modules
 prune changedetectionio/static/styles/package-lock.json
+include changedetectionio/favicon_utils.py
 include changedetection.py
 include requirements.txt
 include README-pip.md

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__ = '0.52.7'
+__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)}")
     
@@ -112,31 +124,85 @@ def sigshutdown_handler(_signo, _stack_frame):
         except Exception as e:
             logger.error(f"Error shutting down Socket.IO server: {str(e)}")
     
-    # Save data quickly
+    # Save data quickly - force immediate save using abstract method
     try:
-        datastore.sync_to_json()
-        logger.success('Fast sync to disk complete.')
+        datastore.force_save_all()
+        logger.success('Fast sync to storage complete.')
     except Exception as e:
-        logger.error(f"Error syncing to disk: {str(e)}")
+        logger.error(f"Error syncing to storage: {str(e)}")
     
     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 +211,68 @@ 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:], "6Ccsd:h:p:l:P:", "port")
+    except getopt.GetoptError as e:
+        print_help()
+        print(f'Error: {e}')
         sys.exit(2)
 
     create_datastore_dir = False
@@ -173,10 +297,6 @@ 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
@@ -188,6 +308,18 @@ 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 +356,16 @@ 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']}'"
@@ -245,15 +382,209 @@ def main():
         logger.critical(str(e))
         return
 
+    # 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
@@ -311,23 +642,43 @@ 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/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
@@ -42,7 +42,7 @@ 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 +50,7 @@ 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}")
@@ -96,6 +96,16 @@ 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
 

changedetectionio/api/Watch.py

@@ -2,10 +2,11 @@ import os
 import threading
 
 from changedetectionio.validate_url import is_safe_valid_url
+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
@@ -65,47 +66,42 @@ 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()
             return "OK", 200
         elif request.args.get('paused', '') == 'unpaused':
-            self.datastore.data['watching'].get(uuid).unpause()
+            watch_obj.unpause()
             return "OK", 200
         if request.args.get('muted', '') == 'muted':
-            self.datastore.data['watching'].get(uuid).mute()
+            watch_obj.mute()
             return "OK", 200
         elif request.args.get('muted', '') == 'unmuted':
-            self.datastore.data['watching'].get(uuid).unmute()
+            watch_obj.unmute()
             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
 
@@ -139,6 +135,16 @@ 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')
@@ -158,58 +164,18 @@ 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)
 
-        # 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
 
@@ -402,16 +368,9 @@ class WatchFavicon(Resource):
 
         favicon_filename = watch.get_favicon_filename()
         if favicon_filename:
-            try:
-                import magic
-                mime = magic.from_file(
-                    os.path.join(watch.watch_data_dir, favicon_filename),
-                    mime=True
-                )
-            except ImportError:
-                # Fallback, no python-magic
-                import mimetypes
-                mime, encoding = mimetypes.guess_type(favicon_filename)
+            # Use cached MIME type detection
+            filepath = os.path.join(watch.watch_data_dir, favicon_filename)
+            mime = get_favicon_mime_type(filepath)
 
             response = make_response(send_from_directory(watch.watch_data_dir, favicon_filename))
             response.headers['Content-type'] = mime
@@ -450,6 +409,16 @@ 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
+
         extras = copy.deepcopy(json_data)
 
         # Because we renamed 'tag' to 'tags' but don't want to change the API (can do this in v2 of the API)
@@ -463,9 +432,19 @@ class CreateWatch(Resource):
         new_uuid = self.datastore.add_watch(url=url, extras=extras, tag=tags)
         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
@@ -500,7 +479,7 @@ 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 = [
@@ -510,7 +489,7 @@ 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)
@@ -522,7 +501,7 @@ 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."""
@@ -532,7 +511,7 @@ 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/blueprint/backups/__init__.py

@@ -27,11 +27,23 @@ 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():
@@ -90,8 +102,8 @@ 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()
+        # Be sure we're written fresh - force immediate save using abstract method
+        datastore.force_save_all()
         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/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'
@@ -20,7 +20,7 @@ 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}))
+        worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
         return redirect(url_for("watchlist.index"))
 
     @login_required

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

@@ -83,7 +83,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
 
                 # 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 +92,7 @@ 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,

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>

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.mark_watch_dirty(uuid)
         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.mark_watch_dirty(uuid)
         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.mark_watch_dirty(uuid)
         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.mark_watch_dirty(uuid)
         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.mark_watch_dirty(uuid)
         if emit_flash:
             flash(gettext("{} watches errors cleared").format(len(uuids)))
 
@@ -109,7 +114,7 @@ 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 +222,14 @@ 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 +239,14 @@ 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 +262,10 @@ 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 +284,7 @@ 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 +294,7 @@ 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 +310,7 @@ 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 +320,7 @@ 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 +349,7 @@ 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 +367,6 @@ 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 +406,25 @@ 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 +197,7 @@ 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])
             flash(gettext("Updated watch - unpaused!") if request.args.get('unpause_on_save') else gettext("Updated watch."))
 
             # Cleanup any browsersteps session for this watch
@@ -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 = {
@@ -371,6 +334,8 @@ 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):
             latest_filename = list(watch.history.keys())[-1]
@@ -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)

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/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>
@@ -69,7 +73,9 @@ 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">
@@ -205,25 +211,26 @@ html[data-darkmode="true"] .watch-tag-list.tag-{{ class_name }} {
                         </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/favicon_utils.py

@@ -0,0 +1,43 @@
+"""
+Favicon utilities for changedetection.io
+Handles favicon MIME type detection with caching
+"""
+
+from functools import lru_cache
+
+
+@lru_cache(maxsize=1000)
+def get_favicon_mime_type(filepath):
+    """
+    Detect MIME type of favicon by reading file content using puremagic.
+    Results are cached to avoid repeatedly reading the same files.
+
+    Args:
+        filepath: Full path to the favicon file
+
+    Returns:
+        MIME type string (e.g., 'image/png')
+    """
+    mime = None
+
+    try:
+        import puremagic
+        with open(filepath, 'rb') as f:
+            content_bytes = f.read(200)  # Read first 200 bytes
+
+        detections = puremagic.magic_string(content_bytes)
+        if detections:
+            mime = detections[0].mime_type
+    except Exception:
+        pass
+
+    # Fallback to mimetypes if puremagic fails
+    if not mime:
+        import mimetypes
+        mime, _ = mimetypes.guess_type(filepath)
+
+    # Final fallback based on extension
+    if not mime:
+        mime = 'image/x-icon' if filepath.endswith('.ico') else 'image/png'
+
+    return mime

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,
@@ -44,6 +44,8 @@ from changedetectionio.api import Watch, WatchHistory, WatchSingleHistory, Watch
 from changedetectionio.api.Search import Search
 from .time_handler import is_within_schedule
 from changedetectionio.languages import get_available_languages, get_language_codes, get_flag_for_locale, get_timeago_locale
+from changedetectionio.favicon_utils import get_favicon_mime_type
+
 IN_PYTEST = "pytest" in sys.modules or "PYTEST_CURRENT_TEST" in os.environ
 
 datastore = None
@@ -69,9 +71,13 @@ socketio_server = None
 CORS(app)
 
 # Super handy for compressing large BrowserSteps responses and others
-FlaskCompress(app)
-app.config['COMPRESS_MIN_SIZE'] = 4096
+# Flask-Compress handles HTTP compression, Socket.IO compression disabled to prevent memory leak
+compress = FlaskCompress()
+app.config['COMPRESS_MIN_SIZE'] = 2096
 app.config['COMPRESS_MIMETYPES'] = ['text/html', 'text/css', 'text/javascript', 'application/json', 'application/javascript', 'image/svg+xml']
+# Use gzip only - smaller memory footprint than zstd/brotli (4-8KB vs 200-500KB contexts)
+app.config['COMPRESS_ALGORITHM'] = ['gzip']
+compress.init_app(app)
 app.config['TEMPLATES_AUTO_RELOAD'] = False
 
 
@@ -88,6 +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
 
 
@@ -181,7 +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):
@@ -192,13 +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'],
@@ -387,7 +401,10 @@ 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
 
@@ -400,11 +417,27 @@ def changedetection_app(config=None, datastore_o=None):
     language_codes = get_language_codes()
 
     def get_locale():
+        # Locale aliases: map browser language codes to translation directory names
+        # This handles cases where browsers send standard codes (e.g., zh-TW)
+        # but our translations use more specific codes (e.g., zh_Hant_TW)
+        locale_aliases = {
+            'zh-TW': 'zh_Hant_TW',  # Traditional Chinese: browser sends zh-TW, we use zh_Hant_TW
+            'zh_TW': 'zh_Hant_TW',  # Also handle underscore variant
+        }
+
         # 1. Try to get locale from session (user explicitly selected)
         if 'locale' in session:
             return session['locale']
+
         # 2. Fall back to Accept-Language header
-        return request.accept_languages.best_match(language_codes)
+        # Get the best match from browser's Accept-Language header
+        browser_locale = request.accept_languages.best_match(language_codes + list(locale_aliases.keys()))
+
+        # 3. Check if we need to map the browser locale to our internal locale
+        if browser_locale in locale_aliases:
+            return locale_aliases[browser_locale]
+
+        return browser_locale
 
     # Initialize Babel with locale selector
     babel = Babel(app, locale_selector=get_locale)
@@ -528,6 +561,9 @@ 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)
@@ -666,16 +702,9 @@ def changedetection_app(config=None, datastore_o=None):
 
             favicon_filename = watch.get_favicon_filename()
             if favicon_filename:
-                try:
-                    import magic
-                    mime = magic.from_file(
-                        os.path.join(watch.watch_data_dir, favicon_filename),
-                        mime=True
-                    )
-                except ImportError:
-                    # Fallback, no python-magic
-                    import mimetypes
-                    mime, encoding = mimetypes.guess_type(favicon_filename)
+                # Use cached MIME type detection
+                filepath = os.path.join(watch.watch_data_dir, favicon_filename)
+                mime = get_favicon_mime_type(filepath)
 
                 response = make_response(send_from_directory(watch.watch_data_dir, favicon_filename))
                 response.headers['Content-type'] = mime
@@ -772,13 +801,15 @@ 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)
+    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
@@ -807,10 +838,10 @@ 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,
@@ -874,16 +905,31 @@ 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
@@ -918,14 +964,14 @@ 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)
@@ -951,7 +997,7 @@ 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:
@@ -992,7 +1038,7 @@ 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,
@@ -1011,7 +1057,7 @@ 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}
@@ -1117,7 +1163,7 @@ 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/Watch.py

@@ -1,3 +1,6 @@
+import gc
+from copy import copy
+
 from blinker import signal
 from changedetectionio.validate_url import is_safe_valid_url
 
@@ -13,15 +16,16 @@ 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 +41,52 @@ 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:
@@ -83,8 +112,15 @@ class model(watch_base):
         self.__datastore_path = kw.get('datastore_path')
         if kw.get('datastore_path'):
             del kw['datastore_path']
-            
+
+        self.__datastore = kw.get('__datastore')
+        if not self.__datastore:
+            raise ValueError("Watch object requires '__datastore' reference - cannot access global settings without it")
+        if kw.get('__datastore'):
+            del kw['__datastore']
+
         super(model, self).__init__(*arg, **kw)
+
         if kw.get('default'):
             self.update(kw['default'])
             del kw['default']
@@ -95,6 +131,95 @@ class model(watch_base):
         # Be sure the cached timestamp is ready
         bump = self.history
 
+    def __deepcopy__(self, memo):
+        """
+        Custom deepcopy that excludes __datastore to prevent memory leaks.
+
+        CRITICAL FIX: Without this, deepcopy(watch) copies the entire datastore
+        (which contains all other watches), causing exponential memory growth.
+        With 100 watches, this creates 10,000 watch objects in memory (100²).
+
+        This is called by:
+        - api/Watch.py:76 (API endpoint)
+        - processors/base.py:26 (EVERY processor run)
+        - store/__init__.py:544 (clone watch)
+        - And 4+ other locations
+        """
+        from copy import deepcopy
+
+        # Create a new instance without calling __init__ (avoids __datastore requirement)
+        cls = self.__class__
+        new_watch = cls.__new__(cls)
+        memo[id(self)] = new_watch
+
+        # Copy the dict data (all the watch settings)
+        for key, value in self.items():
+            new_watch[key] = deepcopy(value, memo)
+
+        # Copy instance attributes EXCEPT the datastore references
+        # These are cached/computed values that need to be preserved
+        new_watch._model__newest_history_key = self._model__newest_history_key
+        new_watch._model__history_n = self._model__history_n
+        new_watch.jitter_seconds = self.jitter_seconds
+
+        # Copy datastore_path (string, safe to copy)
+        new_watch._model__datastore_path = self._model__datastore_path
+
+        # CRITICAL: Share the datastore reference (don't copy it!)
+        # This is safe because we never modify the datastore through the watch
+        new_watch._model__datastore = self._model__datastore
+
+        # Do NOT copy favicon cache - let it be regenerated on demand
+        # This is just a performance cache (prevents repeated glob operations)
+        # and will be rebuilt automatically on first access
+
+        return new_watch
+
+    def __getstate__(self):
+        """
+        Custom pickle serialization that excludes __datastore.
+
+        This handles pickle/unpickle (used by multiprocessing, caching, etc.)
+        and ensures the datastore reference is never serialized.
+        """
+        # Get the dict data
+        state = dict(self)
+
+        # Add the instance attributes we want to preserve
+        state['__watch_metadata__'] = {
+            'newest_history_key': self._model__newest_history_key,
+            'history_n': self._model__history_n,
+            'jitter_seconds': self.jitter_seconds,
+            'datastore_path': self._model__datastore_path,
+        }
+
+        # NOTE: __datastore and _favicon_filename_cache are intentionally excluded
+        # Both will be regenerated/restored as needed
+        return state
+
+    def __setstate__(self, state):
+        """
+        Custom pickle deserialization.
+
+        WARNING: This creates a Watch without a __datastore reference!
+        The caller MUST set watch._model__datastore after unpickling.
+        """
+        # Extract metadata
+        metadata = state.pop('__watch_metadata__', {})
+
+        # Restore dict data
+        self.update(state)
+
+        # Restore instance attributes
+        self._model__newest_history_key = metadata.get('newest_history_key')
+        self._model__history_n = metadata.get('history_n', 0)
+        self.jitter_seconds = metadata.get('jitter_seconds', 0)
+        self._model__datastore_path = metadata.get('datastore_path')
+
+        # __datastore is NOT restored - caller must set it!
+        # _favicon_filename_cache is NOT restored - will regenerate on demand
+        self._model__datastore = None
+
     @property
     def viewed(self):
         # Don't return viewed when last_viewed is 0 and newest_key is 0
@@ -397,16 +522,49 @@ 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.watch_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
+
+        # 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.watch_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 +573,6 @@ 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
@@ -475,6 +632,15 @@ class model(watch_base):
         self.__newest_history_key = timestamp
         self.__history_n += 1
 
+
+        maxlen = (
+                self.get('history_snapshot_max_length')
+                or (self.__datastore and self.__datastore['settings']['application'].get('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
 
@@ -587,6 +753,11 @@ 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 +774,32 @@ 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.*"))
 
         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.

changedetectionio/model/__init__.py

@@ -32,6 +32,7 @@ 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,

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()
 

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/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/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):
@@ -58,7 +64,7 @@ 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))

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)
-            
-            logger.debug(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
+
+                # 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:
-                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)}")
+                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)}: {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()
@@ -240,7 +240,10 @@ def init_socketio(app, datastore):
                         async_mode=async_mode,
                         cors_allowed_origins=cors_origins,  # None means same-origin only
                         logger=strtobool(os.getenv('SOCKETIO_LOGGING', 'False')),
-                        engineio_logger=strtobool(os.getenv('SOCKETIO_LOGGING', 'False')))
+                        engineio_logger=strtobool(os.getenv('SOCKETIO_LOGGING', 'False')),
+                        # Disable WebSocket compression to prevent memory accumulation
+                        # Flask-Compress already handles HTTP response compression
+                        engineio_options={'http_compression': False, 'compression_threshold': 0})
 
     # Set up event handlers
     logger.info("Socket.IO: Registering connect event handler")
@@ -249,7 +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
 
@@ -265,7 +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
@@ -91,8 +95,8 @@ REMOVE_REQUESTS_OLD_SCREENSHOTS=false pytest -vv -s --maxfail=1 tests/test_notif
 # And again with brotli+screenshot attachment
 SNAPSHOT_BROTLI_COMPRESSION_THRESHOLD=5 REMOVE_REQUESTS_OLD_SCREENSHOTS=false pytest -vv -s --maxfail=1 --dist=load tests/test_backend.py tests/test_rss.py tests/test_unique_lines.py tests/test_notification.py  tests/test_access_control.py
 
-# Try high concurrency
-FETCH_WORKERS=50 pytest  tests/test_history_consistency.py -vv -l -s
+# Try high concurrency with aggressive worker restarts
+FETCH_WORKERS=50 WORKER_MAX_RUNTIME=2 WORKER_MAX_JOBS=1 pytest  tests/test_history_consistency.py -vv -l -s
 
 # Check file:// will pickup a file when enabled
 echo "Hello world" > /tmp/test-file.txt

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/search-modal.js

@@ -69,6 +69,19 @@
       }
     });
 
+    // Handle Enter key in search input
+    if (searchInput) {
+      searchInput.addEventListener('keydown', function(e) {
+        if (e.key === 'Enter') {
+          e.preventDefault();
+          if (searchForm) {
+            // Trigger form submission programmatically
+            searchForm.dispatchEvent(new Event('submit'));
+          }
+        }
+      });
+    }
+
     // Handle form submission
     if (searchForm) {
       searchForm.addEventListener('submit', function(e) {
@@ -88,8 +101,8 @@
           params.append('tags', tags);
         }
 
-        // Navigate to search results
-        window.location.href = '?' + params.toString();
+        // Navigate to search results (always redirect to watchlist home)
+        window.location.href = '/?' + params.toString();
       });
     }
   });

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 +1,100 @@
+"""
+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
+
+    @abstractmethod
+    def force_save_all(self):
+        """
+        Force immediate synchronous save of all data to storage.
+
+        This is the abstract method for forcing a complete save.
+        Different backends implement this differently:
+        - File backend: Mark all watches/settings dirty, then save
+        - Redis backend: SAVE command or pipeline flush
+        - SQL backend: COMMIT transaction
+
+        Used by:
+        - Backup creation (ensure everything is saved before backup)
+        - Shutdown (ensure all changes are persisted)
+        - Manual save operations
+        """
+        pass

changedetectionio/store/file_saving_datastore.py

@@ -0,0 +1,898 @@
+"""
+File-based datastore with individual watch persistence and dirty tracking.
+
+This module provides the FileSavingDataStore abstract class that implements:
+- Individual watch.json file persistence
+- Hash-based change detection (only save what changed)
+- Periodic audit scan (catches unmarked changes)
+- Background save thread with batched parallel saves
+- Atomic file writes safe for NFS/NAS
+"""
+
+import glob
+import hashlib
+import json
+import os
+import tempfile
+import time
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from threading import Thread
+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')))
+
+# Save interval configuration: How often the background thread saves dirty items
+# Default 10 seconds - increase for less frequent saves, decrease for more frequent
+DATASTORE_SCAN_DIRTY_SAVE_INTERVAL_SECONDS = int(os.getenv('DATASTORE_SCAN_DIRTY_SAVE_INTERVAL_SECONDS', '10'))
+
+# Rolling audit configuration: Scans a fraction of watches each cycle
+# Default: Run audit every 10s, split into 5 shards
+# Full audit completes every 50s (10s × 5 shards)
+# With 56k watches: 56k / 5 = ~11k watches per cycle (~60ms vs 316ms for all)
+# Handles dynamic watch count - recalculates shard boundaries each cycle
+DATASTORE_AUDIT_INTERVAL_SECONDS = int(os.getenv('DATASTORE_AUDIT_INTERVAL_SECONDS', '10'))
+DATASTORE_AUDIT_SHARDS = int(os.getenv('DATASTORE_AUDIT_SHARDS', '5'))
+
+
+# ============================================================================
+# 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
+
+    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_watch_atomic(watch_dir, uuid, watch_dict):
+    """
+    Save a watch to disk using atomic write pattern.
+
+    Convenience wrapper around save_json_atomic for watches.
+
+    Args:
+        watch_dir: Directory for this watch (e.g., /datastore/{uuid})
+        uuid: Watch UUID (for logging)
+        watch_dict: Dictionary representation of the watch
+
+    Raises:
+        ValueError: If serialized data exceeds 10MB (indicates bug or corruption)
+        OSError: If disk is full (ENOSPC) or other I/O error
+    """
+    watch_json = os.path.join(watch_dir, "watch.json")
+    save_json_atomic(watch_json, watch_dict, label=f"watch {uuid}", 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:
+        Tuple of (Watch object, raw_data_dict) or (None, None) if failed
+        The raw_data_dict is needed to compute the hash before rehydration
+    """
+    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, 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)
+
+        if watch_data.get('time_schedule_limit'):
+            del watch_data['time_schedule_limit']
+        if watch_data.get('time_between_check'):
+            del watch_data['time_between_check']
+
+        # Return both the raw data and the rehydrated watch
+        # Raw data is needed to compute hash before rehydration changes anything
+        watch_obj = rehydrate_entity_func(uuid, watch_data)
+        return watch_obj, watch_data
+
+    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, 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, 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, None
+    except Exception as e:
+        logger.error(f"Failed to load watch {uuid} from {watch_json}: {e}")
+        return None, None
+
+
+def load_all_watches(datastore_path, rehydrate_entity_func, compute_hash_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
+        compute_hash_func: Function to compute hash from raw watch dict
+
+    Returns:
+        Tuple of (watching_dict, hashes_dict)
+        - watching_dict: uuid -> Watch object
+        - hashes_dict: uuid -> hash string (computed from raw data)
+    """
+    start_time = time.time()
+    logger.info("Loading watches from individual watch.json files...")
+
+    watching = {}
+    watch_hashes = {}
+
+    if not os.path.exists(datastore_path):
+        return watching, watch_hashes
+
+    # 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, raw_data = load_watch_from_file(watch_json, uuid_dir, rehydrate_entity_func)
+        if watch and raw_data:
+            watching[uuid_dir] = watch
+            # Compute hash from rehydrated Watch object (as dict) to match how we compute on save
+            # This ensures hash matches what audit will compute from dict(watch)
+            watch_hashes[uuid_dir] = compute_hash_func(dict(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, watch_hashes
+
+
+# ============================================================================
+# FileSavingDataStore Class
+# ============================================================================
+
+class FileSavingDataStore(DataStore):
+    """
+    Abstract datastore that provides file persistence with change tracking.
+
+    Features:
+    - Individual watch.json files (one per watch)
+    - Dirty tracking: Only saves items that have changed
+    - Hash-based change detection: Prevents unnecessary writes
+    - Background save thread: Non-blocking persistence
+    - Two-tier urgency: Standard (60s) and urgent (immediate) saves
+
+    Subclasses must implement:
+    - rehydrate_entity(): Convert dict to Watch object
+    - Access to internal __data structure for watch management
+    """
+
+    needs_write = False
+    needs_write_urgent = False
+    stop_thread = False
+
+    # Change tracking
+    _dirty_watches = set()      # Watch UUIDs that need saving
+    _dirty_settings = False     # Settings changed
+    _watch_hashes = {}          # UUID -> SHA256 hash for change detection
+
+    # Health monitoring
+    _last_save_time = 0         # Timestamp of last successful save
+    _last_audit_time = 0        # Timestamp of last audit scan
+    _save_cycle_count = 0       # Number of save cycles completed
+    _total_saves = 0            # Total watches saved (lifetime)
+    _save_errors = 0            # Total save errors (lifetime)
+    _audit_count = 0            # Number of audit scans completed
+    _audit_found_changes = 0    # Total unmarked changes found by audits
+    _audit_shard_index = 0      # Current shard being audited (rolling audit)
+
+    def __init__(self):
+        super().__init__()
+        self.save_data_thread = None
+        self._last_save_time = time.time()
+        self._last_audit_time = time.time()
+
+
+    def mark_watch_dirty(self, uuid):
+        """
+        Mark a watch as needing save.
+
+        Args:
+            uuid: Watch UUID
+        """
+        with self.lock:
+            self._dirty_watches.add(uuid)
+            dirty_count = len(self._dirty_watches)
+
+        # Backpressure detection - warn if dirty set grows too large
+        if dirty_count > 1000:
+            logger.critical(
+                f"BACKPRESSURE WARNING: {dirty_count} watches pending save! "
+                f"Save thread may not be keeping up with write rate. "
+                f"This could indicate disk I/O bottleneck or save thread failure."
+            )
+        elif dirty_count > 500:
+            logger.warning(
+                f"Dirty watch count high: {dirty_count} watches pending save. "
+                f"Monitoring for potential backpressure."
+            )
+
+        self.needs_write = True
+
+    def mark_settings_dirty(self):
+        """Mark settings as needing save."""
+        with self.lock:
+            self._dirty_settings = True
+        self.needs_write = True
+
+    def _compute_hash(self, watch_dict):
+        """
+        Compute SHA256 hash of watch for change detection.
+
+        Args:
+            watch_dict: Dictionary representation of watch
+
+        Returns:
+            Hex string of SHA256 hash
+        """
+        # Use orjson for deterministic serialization if available
+        if HAS_ORJSON:
+            json_bytes = orjson.dumps(watch_dict, option=orjson.OPT_SORT_KEYS)
+        else:
+            json_str = json.dumps(watch_dict, sort_keys=True, ensure_ascii=False)
+            json_bytes = json_str.encode('utf-8')
+
+        return hashlib.sha256(json_bytes).hexdigest()
+
+    def save_watch(self, uuid, force=False, watch_dict=None, current_hash=None):
+        """
+        Save a single watch if it has changed (polymorphic method).
+
+        Args:
+            uuid: Watch UUID
+            force: If True, skip hash check and save anyway
+            watch_dict: Pre-computed watch dictionary (optimization)
+            current_hash: Pre-computed hash (optimization)
+
+        Returns:
+            True if saved, False if skipped (unchanged)
+        """
+        if not self._watch_exists(uuid):
+            logger.warning(f"Cannot save watch {uuid} - does not exist")
+            return False
+
+        # Get watch dict if not provided
+        if watch_dict is None:
+            watch_dict = self._get_watch_dict(uuid)
+
+        # Compute hash if not provided
+        if current_hash is None:
+            current_hash = self._compute_hash(watch_dict)
+
+        # Skip save if unchanged (unless forced)
+        if not force and current_hash == self._watch_hashes.get(uuid):
+            return False
+
+        try:
+            self._save_watch(uuid, watch_dict)
+            self._watch_hashes[uuid] = current_hash
+            logger.debug(f"Saved watch {uuid}")
+            return True
+        except Exception as e:
+            logger.error(f"Failed to save watch {uuid}: {e}")
+            raise
+
+    def _save_watch(self, uuid, watch_dict):
+        """
+        Save a single watch to storage (polymorphic).
+
+        Backend-specific implementation. Subclasses override for different storage:
+        - File backend: Writes to {uuid}/watch.json
+        - Redis backend: SET watch:{uuid}
+        - SQL backend: UPDATE watches WHERE uuid=?
+
+        Args:
+            uuid: Watch UUID
+            watch_dict: Dictionary representation of watch
+        """
+        # Default file implementation
+        watch_dir = os.path.join(self.datastore_path, uuid)
+        save_watch_atomic(watch_dir, uuid, watch_dict)
+
+    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")
+
+    def _save_dirty_items(self):
+        """
+        Save dirty watches and settings.
+
+        This is the core optimization: instead of saving the entire datastore,
+        we only save watches that were marked dirty and settings if changed.
+        """
+        start_time = time.time()
+
+        # Capture dirty sets under lock
+        with self.lock:
+            dirty_watches = list(self._dirty_watches)
+            dirty_settings = self._dirty_settings
+            self._dirty_watches.clear()
+            self._dirty_settings = False
+
+        if not dirty_watches and not dirty_settings:
+            return
+
+        logger.trace(f"Saving {len(dirty_watches)} dirty watches, settings_dirty={dirty_settings}")
+
+        # Save each dirty watch using the polymorphic save method
+        saved_count = 0
+        error_count = 0
+        skipped_unchanged = 0
+
+        # Process in batches of 50, using thread pool for parallel saves
+        BATCH_SIZE = 50
+        MAX_WORKERS = 20  # Number of parallel save threads
+
+        def save_single_watch(uuid):
+            """Helper function for thread pool execution."""
+            try:
+                # Check if watch still exists (might have been deleted)
+                if not self._watch_exists(uuid):
+                    # Watch was deleted, remove hash
+                    self._watch_hashes.pop(uuid, None)
+                    return {'status': 'deleted', 'uuid': uuid}
+
+                # Pre-check hash to avoid unnecessary save_watch() calls
+                watch_dict = self._get_watch_dict(uuid)
+                current_hash = self._compute_hash(watch_dict)
+
+                if current_hash == self._watch_hashes.get(uuid):
+                    # Watch hasn't actually changed, skip
+                    return {'status': 'unchanged', 'uuid': uuid}
+
+                # Pass pre-computed values to avoid redundant serialization/hashing
+                if self.save_watch(uuid, force=True, watch_dict=watch_dict, current_hash=current_hash):
+                    return {'status': 'saved', 'uuid': uuid}
+                else:
+                    return {'status': 'skipped', 'uuid': uuid}
+            except Exception as e:
+                logger.error(f"Error saving watch {uuid}: {e}")
+                return {'status': 'error', 'uuid': uuid, 'error': e}
+
+        # Process dirty watches in batches
+        for batch_start in range(0, len(dirty_watches), BATCH_SIZE):
+            batch = dirty_watches[batch_start:batch_start + BATCH_SIZE]
+            batch_num = (batch_start // BATCH_SIZE) + 1
+            total_batches = (len(dirty_watches) + BATCH_SIZE - 1) // BATCH_SIZE
+
+            if len(dirty_watches) > BATCH_SIZE:
+                logger.trace(f"Save batch {batch_num}/{total_batches} ({len(batch)} watches)")
+
+            # Use thread pool to save watches in parallel
+            with ThreadPoolExecutor(max_workers=MAX_WORKERS) as executor:
+                # Submit all save tasks
+                future_to_uuid = {executor.submit(save_single_watch, uuid): uuid for uuid in batch}
+
+                # Collect results as they complete
+                for future in as_completed(future_to_uuid):
+                    result = future.result()
+                    status = result['status']
+
+                    if status == 'saved':
+                        saved_count += 1
+                    elif status == 'unchanged':
+                        skipped_unchanged += 1
+                    elif status == 'error':
+                        error_count += 1
+                        # Re-mark for retry
+                        with self.lock:
+                            self._dirty_watches.add(result['uuid'])
+                    # 'deleted' and 'skipped' don't need special handling
+
+        # Save settings if changed
+        if dirty_settings:
+            try:
+                self._save_settings()
+                logger.debug("Saved settings")
+            except Exception as e:
+                logger.error(f"Failed to save settings: {e}")
+                error_count += 1
+                with self.lock:
+                    self._dirty_settings = True
+
+        # Update metrics
+        elapsed = time.time() - start_time
+        self._save_cycle_count += 1
+        self._total_saves += saved_count
+        self._save_errors += error_count
+        self._last_save_time = time.time()
+
+        # Log performance metrics
+        if saved_count > 0:
+            avg_time_per_watch = (elapsed / saved_count) * 1000  # milliseconds
+            skipped_msg = f", {skipped_unchanged} unchanged" if skipped_unchanged > 0 else ""
+            parallel_msg = f" [parallel: {MAX_WORKERS} workers]" if saved_count > 1 else ""
+            logger.info(
+                f"Successfully saved {saved_count} watches in {elapsed:.2f}s "
+                f"(avg {avg_time_per_watch:.1f}ms per watch{skipped_msg}){parallel_msg}. "
+                f"Total: {self._total_saves} saves, {self._save_errors} errors (lifetime)"
+            )
+        elif skipped_unchanged > 0:
+            logger.debug(f"Save cycle: {skipped_unchanged} watches verified unchanged (hash match), nothing saved")
+
+        if error_count > 0:
+            logger.error(f"Save cycle completed with {error_count} errors")
+
+        self.needs_write = False
+        self.needs_write_urgent = False
+
+    def _watch_exists(self, uuid):
+        """
+        Check if watch exists. Subclass must implement.
+
+        Args:
+            uuid: Watch UUID
+
+        Returns:
+            bool
+        """
+        raise NotImplementedError("Subclass must implement _watch_exists")
+
+    def _get_watch_dict(self, uuid):
+        """
+        Get watch as dictionary. Subclass must implement.
+
+        Args:
+            uuid: Watch UUID
+
+        Returns:
+            Dictionary representation of watch
+        """
+        raise NotImplementedError("Subclass must implement _get_watch_dict")
+
+    def _audit_all_watches(self):
+        """
+        Rolling audit: Scans a fraction of watches to detect unmarked changes.
+
+        Instead of scanning ALL watches at once, this scans 1/N shards per cycle.
+        The shard rotates each cycle, completing a full audit every N cycles.
+
+        Handles dynamic watch count - recalculates shard boundaries each cycle,
+        so newly added watches will be audited in subsequent cycles.
+
+        Benefits:
+        - Lower CPU per cycle (56k / 5 = ~11k watches vs all 56k)
+        - More frequent audits overall (every 50s vs every 10s)
+        - Spreads load evenly across time
+        """
+        audit_start = time.time()
+
+        # Get list of all watch UUIDs (read-only, no lock needed)
+        try:
+            all_uuids = list(self.data['watching'].keys())
+        except (KeyError, AttributeError, RuntimeError):
+            # Data structure not ready or being modified
+            return
+
+        if not all_uuids:
+            return
+
+        total_watches = len(all_uuids)
+
+        # Calculate this cycle's shard boundaries
+        # Example: 56,278 watches / 5 shards = 11,255 watches per shard
+        # Shard 0: [0:11255], Shard 1: [11255:22510], etc.
+        shard_size = (total_watches + DATASTORE_AUDIT_SHARDS - 1) // DATASTORE_AUDIT_SHARDS
+        start_idx = self._audit_shard_index * shard_size
+        end_idx = min(start_idx + shard_size, total_watches)
+
+        # Handle wrap-around (shouldn't happen normally, but defensive)
+        if start_idx >= total_watches:
+            self._audit_shard_index = 0
+            start_idx = 0
+            end_idx = min(shard_size, total_watches)
+
+        # Audit only this shard's watches
+        shard_uuids = all_uuids[start_idx:end_idx]
+
+        changes_found = 0
+        errors = 0
+
+        for uuid in shard_uuids:
+            try:
+                # Get current watch dict and compute hash
+                watch_dict = self._get_watch_dict(uuid)
+                current_hash = self._compute_hash(watch_dict)
+                stored_hash = self._watch_hashes.get(uuid)
+
+                # If hash changed and not already marked dirty, mark it
+                if current_hash != stored_hash:
+                    with self.lock:
+                        if uuid not in self._dirty_watches:
+                            self._dirty_watches.add(uuid)
+                            changes_found += 1
+                            logger.warning(
+                                f"Audit detected unmarked change in watch {uuid[:8]}... current {current_hash:8} stored hash {stored_hash[:8]}"
+                                f"(hash changed but not marked dirty)"
+                            )
+                            self.needs_write = True
+            except Exception as e:
+                errors += 1
+                logger.trace(f"Audit error for watch {uuid[:8]}...: {e}")
+
+        audit_elapsed = (time.time() - audit_start) * 1000  # milliseconds
+
+        # Advance to next shard (wrap around after last shard)
+        self._audit_shard_index = (self._audit_shard_index + 1) % DATASTORE_AUDIT_SHARDS
+
+        # Update metrics
+        self._audit_count += 1
+        self._audit_found_changes += changes_found
+        self._last_audit_time = time.time()
+
+        if changes_found > 0:
+            logger.warning(
+                f"Audit shard {self._audit_shard_index}/{DATASTORE_AUDIT_SHARDS} found {changes_found} "
+                f"unmarked changes in {len(shard_uuids)}/{total_watches} watches ({audit_elapsed:.1f}ms)"
+            )
+        else:
+            logger.trace(
+                f"Audit shard {self._audit_shard_index}/{DATASTORE_AUDIT_SHARDS}: "
+                f"{len(shard_uuids)}/{total_watches} watches checked, 0 changes ({audit_elapsed:.1f}ms)"
+            )
+
+    def save_datastore(self):
+        """
+        Background thread that periodically saves dirty items and audits watches.
+
+        Runs two independent cycles:
+        1. Save dirty items every DATASTORE_SCAN_DIRTY_SAVE_INTERVAL_SECONDS (default 10s)
+        2. Rolling audit: every DATASTORE_AUDIT_INTERVAL_SECONDS (default 10s)
+           - Scans 1/DATASTORE_AUDIT_SHARDS watches per cycle (default 1/5)
+           - Full audit completes every 50s (10s × 5 shards)
+           - Automatically handles new/deleted watches
+
+        Uses 0.5s sleep intervals for responsiveness to urgent saves.
+        """
+        while True:
+            if self.stop_thread:
+                # Graceful shutdown: flush any remaining dirty items before stopping
+                if self.needs_write or self._dirty_watches or self._dirty_settings:
+                    logger.warning("Datastore save thread stopping - flushing remaining dirty items...")
+                    try:
+                        self._save_dirty_items()
+                        logger.info("Graceful shutdown complete - all data saved")
+                    except Exception as e:
+                        logger.critical(f"FAILED to save dirty items during shutdown: {e}")
+                else:
+                    logger.info("Datastore save thread stopping - no dirty items")
+                return
+
+            # Check if it's time to run audit scan (every N seconds)
+            if time.time() - self._last_audit_time >= DATASTORE_AUDIT_INTERVAL_SECONDS:
+                try:
+                    self._audit_all_watches()
+                except Exception as e:
+                    logger.error(f"Error in audit cycle: {e}")
+
+            # Save dirty items if needed
+            if self.needs_write or self.needs_write_urgent:
+                try:
+                    self._save_dirty_items()
+                except Exception as e:
+                    logger.error(f"Error in save cycle: {e}")
+
+            # Timer with early break for urgent saves
+            # Each iteration is 0.5 seconds, so iterations = DATASTORE_SCAN_DIRTY_SAVE_INTERVAL_SECONDS * 2
+            for i in range(DATASTORE_SCAN_DIRTY_SAVE_INTERVAL_SECONDS * 2):
+                time.sleep(0.5)
+                if self.stop_thread or self.needs_write_urgent:
+                    break
+
+    def start_save_thread(self):
+        """Start the background save thread."""
+        if not self.save_data_thread or not self.save_data_thread.is_alive():
+            self.save_data_thread = Thread(target=self.save_datastore, daemon=True, name="DatastoreSaver")
+            self.save_data_thread.start()
+            logger.info("Datastore save thread started")
+
+    def force_save_all(self):
+        """
+        Force immediate synchronous save of all changes to storage.
+
+        File backend implementation of the abstract force_save_all() method.
+        Marks all watches and settings as dirty, then saves immediately.
+
+        Used by:
+        - Backup creation (ensure everything is saved before backup)
+        - Shutdown (ensure all changes are persisted)
+        - Manual save operations
+        """
+        logger.info("Force saving all data to storage...")
+
+        # Mark everything as dirty to ensure complete save
+        for uuid in self.data['watching'].keys():
+            self.mark_watch_dirty(uuid)
+        self.mark_settings_dirty()
+
+        # Save immediately (synchronous)
+        self._save_dirty_items()
+
+        logger.success("All data saved to storage")
+
+    def get_health_status(self):
+        """
+        Get datastore health status for monitoring.
+
+        Returns:
+            dict with health metrics and status
+        """
+        now = time.time()
+        time_since_last_save = now - self._last_save_time
+
+        with self.lock:
+            dirty_count = len(self._dirty_watches)
+
+        is_thread_alive = self.save_data_thread and self.save_data_thread.is_alive()
+
+        # Determine health status
+        if not is_thread_alive:
+            status = "CRITICAL"
+            message = "Save thread is DEAD"
+        elif time_since_last_save > 300:  # 5 minutes
+            status = "WARNING"
+            message = f"No save activity for {time_since_last_save:.0f}s"
+        elif dirty_count > 1000:
+            status = "WARNING"
+            message = f"High backpressure: {dirty_count} watches pending"
+        elif self._save_errors > 0 and (self._save_errors / max(self._total_saves, 1)) > 0.01:
+            status = "WARNING"
+            message = f"High error rate: {self._save_errors} errors"
+        else:
+            status = "HEALTHY"
+            message = "Operating normally"
+
+        return {
+            "status": status,
+            "message": message,
+            "thread_alive": is_thread_alive,
+            "dirty_watches": dirty_count,
+            "dirty_settings": self._dirty_settings,
+            "last_save_seconds_ago": int(time_since_last_save),
+            "save_cycles": self._save_cycle_count,
+            "total_saves": self._total_saves,
+            "total_errors": self._save_errors,
+            "error_rate_percent": round((self._save_errors / max(self._total_saves, 1)) * 100, 2)
+        }

changedetectionio/store/legacy_loader.py

@@ -0,0 +1,66 @@
+"""
+Legacy format loader for url-watches.json.
+
+Provides functions to detect and load from the legacy monolithic JSON format.
+Used during migration (update_26) to transition to individual watch.json files.
+"""
+
+import os
+import json
+from loguru import logger
+
+# Try to import orjson for faster JSON serialization
+try:
+    import orjson
+    HAS_ORJSON = True
+except ImportError:
+    HAS_ORJSON = False
+
+
+def has_legacy_datastore(datastore_path):
+    """
+    Check if a legacy url-watches.json file exists.
+
+    This is used by update_26 to determine if migration is needed.
+
+    Args:
+        datastore_path: Path to datastore directory
+
+    Returns:
+        bool: True if url-watches.json exists
+    """
+    url_watches_json = os.path.join(datastore_path, "url-watches.json")
+    return os.path.exists(url_watches_json)
+
+
+def load_legacy_format(json_store_path):
+    """
+    Load datastore from legacy url-watches.json format.
+
+    Args:
+        json_store_path: Full path to url-watches.json file
+
+    Returns:
+        dict: Loaded datastore data with 'watching', 'settings', etc.
+        None: If file doesn't exist or loading failed
+    """
+    logger.info(f"Loading from legacy format: {json_store_path}")
+
+    if not os.path.isfile(json_store_path):
+        logger.warning(f"Legacy file not found: {json_store_path}")
+        return None
+
+    try:
+        if HAS_ORJSON:
+            with open(json_store_path, 'rb') as f:
+                data = orjson.loads(f.read())
+        else:
+            with open(json_store_path, 'r', encoding='utf-8') as f:
+                data = json.load(f)
+
+        logger.info(f"Loaded {len(data.get('watching', {}))} watches from legacy format")
+        return data
+
+    except Exception as e:
+        logger.error(f"Failed to load legacy format: {e}")
+        return None

changedetectionio/store/updates.py

@@ -0,0 +1,686 @@
+"""
+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
+
+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
+from .file_saving_datastore import save_watch_atomic
+
+
+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
+    - 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 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 directories with their watch.json files
+            # This preserves the UUID directory structure
+            watch_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
+
+                # Check if this directory has a watch.json (indicates it's a watch UUID directory)
+                watch_json = os.path.join(entry_path, "watch.json")
+                if os.path.isfile(watch_json):
+                    # Add the watch.json file preserving directory structure
+                    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...")
+
+            logger.success(f"Backup created: {backup_filename} ({watch_count} watches)")
+            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):
+        """
+        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
+           - Mark settings and watches dirty
+        3. If any update fails, stop processing
+        4. Save all changes immediately
+        """
+        updates_available = self.get_updates_available()
+
+        # 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.mark_settings_dirty()
+                    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.error(f"Error while trying update_{update_n}")
+                    logger.error(e)
+                    # Don't run any more updates
+                    return
+                else:
+                    # Bump the version, important
+                    self.data['settings']['application']['schema_version'] = update_n
+                    self.mark_settings_dirty()
+
+                    # CRITICAL: Mark all watches as dirty so changes are persisted
+                    # Most updates modify watches, and in the new individual watch.json structure,
+                    # we need to ensure those changes are saved
+                    logger.info(f"Marking all {len(self.data['watching'])} watches as dirty after update_{update_n} (so that it saves them to disk)")
+                    for uuid in self.data['watching'].keys():
+                        self.mark_watch_dirty(uuid)
+
+                    # Save changes immediately after each update (more resilient than batching)
+                    logger.critical(f"Saving all changes after update_{update_n}")
+                    try:
+                        self._save_dirty_items()
+                        logger.success(f"Update {update_n} changes saved successfully")
+                    except Exception as e:
+                        logger.error(f"Failed to save update_{update_n} changes: {e}")
+                        # Don't raise - update already ran, but changes might not be persisted
+                        # The update will try to run again on next startup
+
+                    # 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'):
+            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)
+
+        # Check if already migrated
+        changedetection_json = os.path.join(self.datastore_path, "changedetection.json")
+        if os.path.exists(changedetection_json):
+            logger.info("Migration already completed (changedetection.json exists), skipping")
+            return
+
+        # Check if we need to load legacy data
+        from .legacy_loader import has_legacy_datastore, load_legacy_format
+
+        if not has_legacy_datastore(self.datastore_path):
+            logger.info("No legacy datastore found, nothing to migrate")
+            return
+
+        # Load legacy data from url-watches.json
+        logger.critical("Loading legacy datastore from url-watches.json...")
+        legacy_path = os.path.join(self.datastore_path, "url-watches.json")
+        legacy_data = load_legacy_format(legacy_path)
+
+        if not legacy_data:
+            raise Exception("Failed to load legacy datastore from url-watches.json")
+
+        # 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_dict = dict(watch)
+                watch_dir = os.path.join(self.datastore_path, uuid)
+                save_watch_atomic(watch_dir, uuid, watch_dict)
+                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...")
+
+        if not os.path.isfile(changedetection_json):
+            raise Exception(
+                "Migration failed: changedetection.json not found after save. "
+                "url-watches.json remains intact, safe to retry."
+            )
+
+        logger.critical("Phase 4 complete: Verified changedetection.json exists")
+
+        # Success! Now reload from new format
+        logger.critical("Reloading datastore from new format...")
+        self._load_state() # Includes load_watches
+        logger.success("Datastore reloaded from new format successfully")
+
+
+        # Verify all watches have hashes after migration
+        missing_hashes = [uuid for uuid in self.data['watching'].keys() if uuid not in self._watch_hashes]
+        if missing_hashes:
+            logger.error(f"WARNING: {len(missing_hashes)} watches missing hashes after migration: {missing_hashes[:5]}")
+        else:
+            logger.success(f"All {len(self.data['watching'])} watches have valid hashes after migration")
+
+        # Set schema version to latest available update
+        # This prevents re-running updates and re-marking all watches as dirty
+        updates_available = self.get_updates_available()
+        latest_schema = updates_available[-1] if updates_available else 26
+        self.data['settings']['application']['schema_version'] = latest_schema
+        self.mark_settings_dirty()
+        logger.info(f"Set schema_version to {latest_schema} (migration complete, all watches already saved)")
+
+        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()

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/templates/menu.html

@@ -50,7 +50,9 @@
             <span class="visually-hidden">{{ _('Change language') }}</span>
             <span class="{{ get_flag_for_locale(get_locale()) }}" id="language-selector-flag"></span>
         </button>
-        <a class="github-link" href="https://github.com/dgtlmoon/changedetection.io">
+        <a class="github-link" href="https://github.com/dgtlmoon/changedetection.io"
+           target="_blank" 
+          rel="noopener" >
             {% include "svgs/github.svg" %}
         </a>
     </li>

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."""
@@ -167,6 +257,57 @@ def prepare_test_function(live_server, datastore_path):
         except:
             break
 
+    # 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
+
+
+
+
     # Prevent background thread from writing during cleanup/reload
     datastore.needs_write = False
     datastore.needs_write_urgent = False
@@ -264,8 +405,8 @@ 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 +454,3 @@ 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_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_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_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

@@ -166,7 +166,8 @@ 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)

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):
     """
@@ -325,3 +353,274 @@ def test_time_unit_translations(client, live_server, measure_memory_usage, datas
     assert b"Time Between Check" not in res.data, "Should not have English 'Time Between Check'"
     assert "Chrome 請求".encode() not in res.data, "Should not have incorrect 'Chrome 請求' (Chrome requests)"
     assert "使用預設通知".encode() not in res.data, "Should not have incorrect '使用預設通知' (Use default notification)"
+
+
+def test_accept_language_header_zh_tw(client, live_server, measure_memory_usage, datastore_path):
+    """
+    Test that browsers sending zh-TW in Accept-Language header get Traditional Chinese.
+    This tests the locale alias mapping for issue #3779.
+    """
+    from flask import url_for
+
+    # Clear any session data to simulate a fresh visitor
+    with client.session_transaction() as sess:
+        sess.clear()
+
+    # Request the index page with zh-TW in Accept-Language header (what browsers send)
+    res = client.get(
+        url_for("watchlist.index"),
+        headers={'Accept-Language': 'zh-TW,zh;q=0.9,en;q=0.8'},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Should get Traditional Chinese content, not Simplified Chinese
+    # Traditional: 選擇語言, Simplified: 选择语言
+    assert '選擇語言'.encode() in res.data, "Expected Traditional Chinese '選擇語言' (Select Language)"
+    assert '选择语言'.encode() not in res.data, "Should not get Simplified Chinese '选择语言'"
+
+    # Check HTML lang attribute uses BCP 47 format
+    assert b'<html lang="zh-Hant-TW"' in res.data, "Expected BCP 47 language tag zh-Hant-TW in HTML"
+
+    # Check that the correct flag icon is shown (Taiwan flag for Traditional Chinese)
+    assert b'<span class="fi fi-tw fis" id="language-selector-flag">' in res.data, \
+        "Expected Taiwan flag 'fi fi-tw' for Traditional Chinese"
+    assert b'<span class="fi fi-cn fis" id="language-selector-flag">' not in res.data, \
+        "Should not show China flag 'fi fi-cn' for Traditional Chinese"
+
+    # Verify we're getting Traditional Chinese text throughout the page
+    res = client.get(
+        url_for("settings.settings_page"),
+        headers={'Accept-Language': 'zh-TW,zh;q=0.9,en;q=0.8'},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Check Traditional Chinese translations (not English)
+    assert "小時".encode() in res.data, "Expected Traditional Chinese '小時' for Hours"
+    assert "分鐘".encode() in res.data, "Expected Traditional Chinese '分鐘' for Minutes"
+    assert b"Hours" not in res.data or "小時".encode() in res.data, "Should have Traditional Chinese, not English"
+
+
+def test_accept_language_header_en_variants(client, live_server, measure_memory_usage, datastore_path):
+    """
+    Test that browsers sending en-GB and en-US in Accept-Language header get the correct English variant.
+    This ensures the locale selector works properly for English variants.
+    """
+    from flask import url_for
+
+    # Test 1: British English (en-GB)
+    with client.session_transaction() as sess:
+        sess.clear()
+
+    res = client.get(
+        url_for("watchlist.index"),
+        headers={'Accept-Language': 'en-GB,en;q=0.9'},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Should get English content
+    assert b"Select Language" in res.data, "Expected English text 'Select Language'"
+
+    # Check HTML lang attribute uses BCP 47 format with hyphen
+    assert b'<html lang="en-GB"' in res.data, "Expected BCP 47 language tag en-GB in HTML"
+
+    # Check that the correct flag icon is shown (UK flag for en-GB)
+    assert b'<span class="fi fi-gb fis" id="language-selector-flag">' in res.data, \
+        "Expected UK flag 'fi fi-gb' for British English"
+
+    # Test 2: American English (en-US)
+    with client.session_transaction() as sess:
+        sess.clear()
+
+    res = client.get(
+        url_for("watchlist.index"),
+        headers={'Accept-Language': 'en-US,en;q=0.9'},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Should get English content
+    assert b"Select Language" in res.data, "Expected English text 'Select Language'"
+
+    # Check HTML lang attribute uses BCP 47 format with hyphen
+    assert b'<html lang="en-US"' in res.data, "Expected BCP 47 language tag en-US in HTML"
+
+    # Check that the correct flag icon is shown (US flag for en-US)
+    assert b'<span class="fi fi-us fis" id="language-selector-flag">' in res.data, \
+        "Expected US flag 'fi fi-us' for American English"
+
+    # Test 3: Generic 'en' should fall back to one of the English variants
+    with client.session_transaction() as sess:
+        sess.clear()
+
+    res = client.get(
+        url_for("watchlist.index"),
+        headers={'Accept-Language': 'en'},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Should get English content (either variant is fine)
+    assert b"Select Language" in res.data, "Expected English text 'Select Language'"
+
+
+def test_accept_language_header_zh_simplified(client, live_server, measure_memory_usage, datastore_path):
+    """
+    Test that browsers sending zh or zh-CN in Accept-Language header get Simplified Chinese.
+    This ensures Simplified Chinese still works correctly and doesn't get confused with Traditional.
+    """
+    from flask import url_for
+
+    # Test 1: Generic 'zh' should get Simplified Chinese
+    with client.session_transaction() as sess:
+        sess.clear()
+
+    res = client.get(
+        url_for("watchlist.index"),
+        headers={'Accept-Language': 'zh,en;q=0.9'},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Should get Simplified Chinese content, not Traditional Chinese
+    # Simplified: 选择语言, Traditional: 選擇語言
+    assert '选择语言'.encode() in res.data, "Expected Simplified Chinese '选择语言' (Select Language)"
+    assert '選擇語言'.encode() not in res.data, "Should not get Traditional Chinese '選擇語言'"
+
+    # Check HTML lang attribute
+    assert b'<html lang="zh"' in res.data, "Expected language tag zh in HTML"
+
+    # Check that the correct flag icon is shown (China flag for Simplified Chinese)
+    assert b'<span class="fi fi-cn fis" id="language-selector-flag">' in res.data, \
+        "Expected China flag 'fi fi-cn' for Simplified Chinese"
+    assert b'<span class="fi fi-tw fis" id="language-selector-flag">' not in res.data, \
+        "Should not show Taiwan flag 'fi fi-tw' for Simplified Chinese"
+
+    # Test 2: 'zh-CN' should also get Simplified Chinese
+    with client.session_transaction() as sess:
+        sess.clear()
+
+    res = client.get(
+        url_for("watchlist.index"),
+        headers={'Accept-Language': 'zh-CN,zh;q=0.9,en;q=0.8'},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Should get Simplified Chinese content
+    assert '选择语言'.encode() in res.data, "Expected Simplified Chinese '选择语言' with zh-CN header"
+    assert '選擇語言'.encode() not in res.data, "Should not get Traditional Chinese with zh-CN header"
+
+    # Check that the correct flag icon is shown (China flag for zh-CN)
+    assert b'<span class="fi fi-cn fis" id="language-selector-flag">' in res.data, \
+        "Expected China flag 'fi fi-cn' for zh-CN header"
+
+    # Verify Simplified Chinese in settings page
+    res = client.get(
+        url_for("settings.settings_page"),
+        headers={'Accept-Language': 'zh,en;q=0.9'},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Check Simplified Chinese translations (not Traditional or English)
+    # Simplified: 小时, Traditional: 小時
+    assert "小时".encode() in res.data, "Expected Simplified Chinese '小时' for Hours"
+    assert "分钟".encode() in res.data, "Expected Simplified Chinese '分钟' for Minutes"
+    assert "秒".encode() in res.data, "Expected Simplified Chinese '秒' for Seconds"
+    # Make sure it's not Traditional Chinese
+    assert "小時".encode() not in res.data, "Should not have Traditional Chinese '小時'"
+    assert "分鐘".encode() not in res.data, "Should not have Traditional Chinese '分鐘'"
+
+
+def test_session_locale_overrides_accept_language(client, live_server, measure_memory_usage, datastore_path):
+    """
+    Test that session locale preference overrides browser Accept-Language header.
+
+    Scenario:
+    1. Browser auto-detects zh-TW (Traditional Chinese) from Accept-Language header
+    2. User explicitly selects Korean language
+    3. On subsequent page loads, Korean should be shown (not Traditional Chinese)
+       even though the Accept-Language header still says zh-TW
+
+    This tests the session override behavior for issue #3779.
+    """
+    from flask import url_for
+
+    # Step 1: Clear session and make first request with zh-TW header (auto-detect)
+    with client.session_transaction() as sess:
+        sess.clear()
+
+    res = client.get(
+        url_for("watchlist.index"),
+        headers={'Accept-Language': 'zh-TW,zh;q=0.9,en;q=0.8'},
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Should initially get Traditional Chinese from auto-detect
+    assert '選擇語言'.encode() in res.data, "Expected Traditional Chinese '選擇語言' from auto-detect"
+    assert b'<html lang="zh-Hant-TW"' in res.data, "Expected zh-Hant-TW language tag"
+    assert b'<span class="fi fi-tw fis" id="language-selector-flag">' in res.data, \
+        "Expected Taiwan flag 'fi fi-tw' from auto-detect"
+
+    # Step 2: User explicitly selects Korean language
+    res = client.get(
+        url_for("set_language", locale="ko"),
+        headers={'Accept-Language': 'zh-TW,zh;q=0.9,en;q=0.8'},  # Browser still sends zh-TW
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Step 3: Make another request with same zh-TW header
+    # Session should override the Accept-Language header
+    res = client.get(
+        url_for("watchlist.index"),
+        headers={'Accept-Language': 'zh-TW,zh;q=0.9,en;q=0.8'},  # Still sending zh-TW!
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Should now get Korean (session overrides auto-detect)
+    # Korean: 언어 선택, Traditional Chinese: 選擇語言
+    assert '언어 선택'.encode() in res.data, "Expected Korean '언어 선택' (Select Language) from session"
+    assert '選擇語言'.encode() not in res.data, "Should not get Traditional Chinese when Korean is set in session"
+
+    # Check HTML lang attribute is Korean
+    assert b'<html lang="ko"' in res.data, "Expected Korean language tag 'ko' in HTML"
+
+    # Check that Korean flag is shown (not Taiwan flag)
+    assert b'<span class="fi fi-kr fis" id="language-selector-flag">' in res.data, \
+        "Expected Korean flag 'fi fi-kr' from session preference"
+    assert b'<span class="fi fi-tw fis" id="language-selector-flag">' not in res.data, \
+        "Should not show Taiwan flag when Korean is set in session"
+
+    # Verify Korean text on settings page as well
+    res = client.get(
+        url_for("settings.settings_page"),
+        headers={'Accept-Language': 'zh-TW,zh;q=0.9,en;q=0.8'},  # Still zh-TW!
+        follow_redirects=True
+    )
+
+    assert res.status_code == 200
+
+    # Check Korean translations (not Traditional Chinese or English)
+    # Korean: 시간 (Hours), 분 (Minutes), 초 (Seconds)
+    # Traditional Chinese: 小時, 分鐘, 秒
+    assert "시간".encode() in res.data, "Expected Korean '시간' for Hours"
+    assert "분".encode() in res.data, "Expected Korean '분' for Minutes"
+    assert "小時".encode() not in res.data, "Should not have Traditional Chinese '小時' when Korean is set"
+    assert "分鐘".encode() not in res.data, "Should not have Traditional Chinese '分鐘' when Korean is set"

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

@@ -13,7 +13,14 @@ 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 +56,7 @@ 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"
 

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 +160,40 @@ def extract_UUID_from_client(client):
     return uuid.strip()
 
 def delete_all_watches(client=None):
+
+    # Change tracking
+    client.application.config.get('DATASTORE')._dirty_watches = set()      # Watch UUIDs that need saving
+    client.application.config.get('DATASTORE')._dirty_settings = False     # Settings changed
+    client.application.config.get('DATASTORE')._watch_hashes = {}          # UUID -> SHA256 hash for change detection
+
     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 +242,11 @@ 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 +268,35 @@ 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 +371,12 @@ 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