adding notes

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

@@ -395,6 +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

Dockerfile

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

changedetectionio/__init__.py

@@ -112,9 +112,9 @@ def sigshutdown_handler(_signo, _stack_frame):
         from changedetectionio.flask_app import update_q, notification_q
         update_q.close()
         notification_q.close()
-        logger.debug("Janus queues closed successfully")
+        logger.debug("Queues closed successfully")
     except Exception as e:
-        logger.critical(f"CRITICAL: Failed to close janus queues: {e}")
+        logger.critical(f"CRITICAL: Failed to close queues: {e}")
     
     # Shutdown socketio server fast
     from changedetectionio.flask_app import socketio_server
@@ -124,13 +124,9 @@ def sigshutdown_handler(_signo, _stack_frame):
         except Exception as e:
             logger.error(f"Error shutting down Socket.IO server: {str(e)}")
     
-    # Save data quickly - force immediate save using abstract method
-    try:
-        datastore.force_save_all()
-        logger.success('Fast sync to storage complete.')
-    except Exception as e:
-        logger.error(f"Error syncing to storage: {str(e)}")
-    
+    # With immediate persistence, all data is already saved
+    logger.success('All data already persisted (immediate commits enabled).')
+
     sys.exit()
 
 def print_help():

changedetectionio/api/Notifications.py

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

changedetectionio/api/Tags.py

@@ -63,9 +63,11 @@ class Tag(Resource):
 
         if request.args.get('muted', '') == 'muted':
             self.datastore.data['settings']['application']['tags'][uuid]['notification_muted'] = True
+            self.datastore.commit()
             return "OK", 200
         elif request.args.get('muted', '') == 'unmuted':
             self.datastore.data['settings']['application']['tags'][uuid]['notification_muted'] = False
+            self.datastore.commit()
             return "OK", 200
 
         return tag
@@ -79,11 +81,13 @@ class Tag(Resource):
 
         # Delete the tag, and any tag reference
         del self.datastore.data['settings']['application']['tags'][uuid]
-        
+        self.datastore.commit()
+
         # Remove tag from all watches
         for watch_uuid, watch in self.datastore.data['watching'].items():
             if watch.get('tags') and uuid in watch['tags']:
                 watch['tags'].remove(uuid)
+                watch.commit()
 
         return 'OK', 204
 
@@ -107,7 +111,7 @@ class Tag(Resource):
                 return str(e), 400
 
         tag.update(request.json)
-        self.datastore.needs_write_urgent = True
+        self.datastore.commit()
 
         return "OK", 200
 

changedetectionio/api/Watch.py

@@ -66,47 +66,46 @@ class Watch(Resource):
     @validate_openapi_request('getWatch')
     def get(self, uuid):
         """Get information about a single watch, recheck, pause, or mute."""
-        import time
-        from copy import deepcopy
-        watch = None
-        # Retry up to 20 times if dict is being modified
-        # With sleep(0), this is fast: ~200µs best case, ~20ms worst case under heavy load
-        for attempt in range(20):
-            try:
-                watch = deepcopy(self.datastore.data['watching'].get(uuid))
-                break
-            except RuntimeError:
-                # Dict changed during deepcopy, retry after yielding to scheduler
-                # sleep(0) releases GIL and yields - no fixed delay, just lets other threads run
-                if attempt < 19:  # Don't yield on last attempt
-                    time.sleep(0)  # Yield to scheduler (microseconds, not milliseconds)
-
-        if not watch:
+        # Get watch reference first (for pause/mute operations)
+        watch_obj = self.datastore.data['watching'].get(uuid)
+        if not watch_obj:
             abort(404, message='No watch exists with the UUID of {}'.format(uuid))
 
+        # Create a dict copy for JSON response (with lock for thread safety)
+        # This is much faster than deepcopy and doesn't copy the datastore reference
+        # WARNING: dict() is a SHALLOW copy - nested dicts are shared with original!
+        # Only safe because we only ADD scalar properties (line 97-101), never modify nested dicts
+        # If you need to modify nested dicts, use: from copy import deepcopy; watch = deepcopy(dict(watch_obj))
+        with self.datastore.lock:
+            watch = dict(watch_obj)
+
         if request.args.get('recheck'):
             worker_pool.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
             return "OK", 200
         if request.args.get('paused', '') == 'paused':
-            self.datastore.data['watching'].get(uuid).pause()
+            watch_obj.pause()
+            watch_obj.commit()
             return "OK", 200
         elif request.args.get('paused', '') == 'unpaused':
-            self.datastore.data['watching'].get(uuid).unpause()
+            watch_obj.unpause()
+            watch_obj.commit()
             return "OK", 200
         if request.args.get('muted', '') == 'muted':
-            self.datastore.data['watching'].get(uuid).mute()
+            watch_obj.mute()
+            watch_obj.commit()
             return "OK", 200
         elif request.args.get('muted', '') == 'unmuted':
-            self.datastore.data['watching'].get(uuid).unmute()
+            watch_obj.unmute()
+            watch_obj.commit()
             return "OK", 200
 
         # Return without history, get that via another API call
         # Properties are not returned as a JSON, so add the required props manually
-        watch['history_n'] = watch.history_n
+        watch['history_n'] = watch_obj.history_n
         # attr .last_changed will check for the last written text snapshot on change
-        watch['last_changed'] = watch.last_changed
-        watch['viewed'] = watch.viewed
-        watch['link'] = watch.link,
+        watch['last_changed'] = watch_obj.last_changed
+        watch['viewed'] = watch_obj.viewed
+        watch['link'] = watch_obj.link,
 
         return watch
 
@@ -169,58 +168,19 @@ class Watch(Resource):
 
         # Handle processor-config-* fields separately (save to JSON, not datastore)
         from changedetectionio import processors
-        processor_config_data = {}
-        regular_data = {}
 
-        for key, value in request.json.items():
-            if key.startswith('processor_config_'):
-                config_key = key.replace('processor_config_', '')
-                if value:  # Only save non-empty values
-                    processor_config_data[config_key] = value
-            else:
-                regular_data[key] = value
+        # Make a mutable copy of request.json for modification
+        json_data = dict(request.json)
+
+        # Extract and remove processor config fields from json_data
+        processor_config_data = processors.extract_processor_config_from_form_data(json_data)
 
         # Update watch with regular (non-processor-config) fields
-        watch.update(regular_data)
+        watch.update(json_data)
+        watch.commit()
 
-        # Save processor config to JSON file if any config data exists
-        if processor_config_data:
-            try:
-                processor_name = request.json.get('processor', watch.get('processor'))
-                if processor_name:
-                    # Create a processor instance to access config methods
-                    from changedetectionio.processors import difference_detection_processor
-                    processor_instance = difference_detection_processor(self.datastore, uuid)
-                    # Use processor name as filename so each processor keeps its own config
-                    config_filename = f'{processor_name}.json'
-                    processor_instance.update_extra_watch_config(config_filename, processor_config_data)
-                    logger.debug(f"API: Saved processor config to {config_filename}: {processor_config_data}")
-
-                    # Call optional edit_hook if processor has one
-                    try:
-                        import importlib
-                        edit_hook_module_name = f'changedetectionio.processors.{processor_name}.edit_hook'
-
-                        try:
-                            edit_hook = importlib.import_module(edit_hook_module_name)
-                            logger.debug(f"API: Found edit_hook module for {processor_name}")
-
-                            if hasattr(edit_hook, 'on_config_save'):
-                                logger.info(f"API: Calling edit_hook.on_config_save for {processor_name}")
-                                # Call hook and get updated config
-                                updated_config = edit_hook.on_config_save(watch, processor_config_data, self.datastore)
-                                # Save updated config back to file
-                                processor_instance.update_extra_watch_config(config_filename, updated_config)
-                                logger.info(f"API: Edit hook updated config: {updated_config}")
-                            else:
-                                logger.debug(f"API: Edit hook module found but no on_config_save function")
-                        except ModuleNotFoundError:
-                            logger.debug(f"API: No edit_hook module for processor {processor_name} (this is normal)")
-                    except Exception as hook_error:
-                        logger.error(f"API: Edit hook error (non-fatal): {hook_error}", exc_info=True)
-
-            except Exception as e:
-                logger.error(f"API: Failed to save processor config: {e}")
+        # Save processor config to JSON file
+        processors.save_processor_config(self.datastore, uuid, processor_config_data)
 
         return "OK", 200
 
@@ -464,8 +424,14 @@ class CreateWatch(Resource):
             except ValidationError as e:
                 return str(e), 400
 
+        # Handle processor-config-* fields separately (save to JSON, not watch)
+        from changedetectionio import processors
+
         extras = copy.deepcopy(json_data)
 
+        # Extract and remove processor config fields from extras
+        processor_config_data = processors.extract_processor_config_from_form_data(extras)
+
         # Because we renamed 'tag' to 'tags' but don't want to change the API (can do this in v2 of the API)
         tags = None
         if extras.get('tag'):
@@ -475,6 +441,10 @@ class CreateWatch(Resource):
         del extras['url']
 
         new_uuid = self.datastore.add_watch(url=url, extras=extras, tag=tags)
+
+        # Save processor config to separate JSON file
+        if new_uuid and processor_config_data:
+            processors.save_processor_config(self.datastore, new_uuid, processor_config_data)
         if new_uuid:
 # Dont queue because the scheduler will check that it hasnt been checked before anyway
 #            worker_pool.queue_item_async_safe(self.update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': new_uuid}))

changedetectionio/api/__init__.py

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

changedetectionio/blueprint/backups/__init__.py

@@ -102,8 +102,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
             flash(gettext("Maximum number of backups reached, please remove some"), "error")
             return redirect(url_for('backups.index'))
 
-        # Be sure we're written fresh - force immediate save using abstract method
-        datastore.force_save_all()
+        # With immediate persistence, all data is already saved
         zip_thread = threading.Thread(
             target=create_backup,
             args=(datastore.datastore_path, datastore.data.get("watching")),

changedetectionio/blueprint/backups/templates/overview.html

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

changedetectionio/blueprint/imports/__init__.py

@@ -26,8 +26,9 @@ 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:

changedetectionio/blueprint/price_data_follower/__init__.py

@@ -20,6 +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()
+        datastore.data['watching'][uuid].commit()
         worker_pool.queue_item_async_safe(update_q, queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid}))
         return redirect(url_for("watchlist.index"))
 
@@ -27,6 +28,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q: PriorityQueue
     @price_data_follower_blueprint.route("/<string:uuid>/reject", methods=['GET'])
     def reject(uuid):
         datastore.data['watching'][uuid]['track_ldjson_price_data'] = PRICE_DATA_TRACK_REJECT
+        datastore.data['watching'][uuid].commit()
         return redirect(url_for("watchlist.index"))
 
 

changedetectionio/blueprint/settings/__init__.py

@@ -74,12 +74,13 @@ def construct_blueprint(datastore: ChangeDetectionStore):
                     del (app_update['password'])
 
                 datastore.data['settings']['application'].update(app_update)
-                
+
                 # Handle dynamic worker count adjustment
                 old_worker_count = datastore.data['settings']['requests'].get('workers', 1)
                 new_worker_count = form.data['requests'].get('workers', 1)
 
                 datastore.data['settings']['requests'].update(form.data['requests'])
+                datastore.commit()
 
                 # Adjust worker count if it changed
                 if new_worker_count != old_worker_count:
@@ -109,13 +110,11 @@ def construct_blueprint(datastore: ChangeDetectionStore):
 
                 if not os.getenv("SALTED_PASS", False) and len(form.application.form.password.encrypted_password):
                     datastore.data['settings']['application']['password'] = form.application.form.password.encrypted_password
-                    datastore.needs_write_urgent = True
+                    datastore.commit()
                     flash(gettext("Password protection enabled."), 'notice')
                     flask_login.logout_user()
                     return redirect(url_for('watchlist.index'))
 
-                datastore.needs_write_urgent = True
-
                 # Also save plugin settings from the same form submission
                 plugin_tabs_list = get_plugin_settings_tabs()
                 for tab in plugin_tabs_list:
@@ -181,7 +180,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     def settings_reset_api_key():
         secret = secrets.token_hex(16)
         datastore.data['settings']['application']['api_access_token'] = secret
-        datastore.needs_write_urgent = True
+        datastore.commit()
         flash(gettext("API Key was regenerated."))
         return redirect(url_for('settings.settings_page')+'#api')
         
@@ -198,7 +197,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     def toggle_all_paused():
         current_state = datastore.data['settings']['application'].get('all_paused', False)
         datastore.data['settings']['application']['all_paused'] = not current_state
-        datastore.needs_write_urgent = True
+        datastore.commit()
 
         if datastore.data['settings']['application']['all_paused']:
             flash(gettext("Automatic scheduling paused - checks will not be queued."), 'notice')
@@ -212,7 +211,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     def toggle_all_muted():
         current_state = datastore.data['settings']['application'].get('all_muted', False)
         datastore.data['settings']['application']['all_muted'] = not current_state
-        datastore.needs_write_urgent = True
+        datastore.commit()
 
         if datastore.data['settings']['application']['all_muted']:
             flash(gettext("All notifications muted."), 'notice')

changedetectionio/blueprint/settings/templates/settings.html

@@ -25,7 +25,7 @@
             <li class="tab"><a href="#ui-options">{{ _('UI Options') }}</a></li>
             <li class="tab"><a href="#api">{{ _('API') }}</a></li>
             <li class="tab"><a href="#rss">{{ _('RSS') }}</a></li>
-            <li class="tab"><a href="{{ url_for('backups.index') }}" class="pure-menu-link">{{ _('Backups') }}</a></li>
+            <li class="tab"><a href="{{ url_for('backups.index') }}">{{ _('Backups') }}</a></li>
             <li class="tab"><a href="#timedate">{{ _('Time & Date') }}</a></li>
             <li class="tab"><a href="#proxies">{{ _('CAPTCHA & Proxies') }}</a></li>
             {% if plugin_tabs %}
@@ -59,6 +59,14 @@
                         {{ _('Set to') }} <strong>0</strong> {{ _('to disable') }}
                         </span>
                     </div>
+                    <div class="pure-control-group">
+                        {{ render_field(form.application.form.history_snapshot_max_length, class="history_snapshot_max_length") }}
+                        <span class="pure-form-message-inline">{{ _('Limit collection of history snapshots for each watch to this number of history items.') }}
+                            <br>
+                        {{ _('Set to empty to disable / no limit') }}
+                        </span>
+                    </div>
+
                     <div class="pure-control-group">
                         {% if not hide_remove_pass %}
                             {% if current_user.is_authenticated %}

changedetectionio/blueprint/tags/__init__.py

@@ -59,6 +59,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     def mute(uuid):
         if datastore.data['settings']['application']['tags'].get(uuid):
             datastore.data['settings']['application']['tags'][uuid]['notification_muted'] = not datastore.data['settings']['application']['tags'][uuid]['notification_muted']
+            datastore.commit()
         return redirect(url_for('tags.tags_overview_page'))
 
     @tags_blueprint.route("/delete/<string:uuid>", methods=['GET'])
@@ -76,6 +77,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
                 for watch_uuid, watch in datastore.data['watching'].items():
                     if watch.get('tags') and tag_uuid in watch['tags']:
                         watch['tags'].remove(tag_uuid)
+                        watch.commit()
                         removed_count += 1
                 logger.info(f"Background: Tag {tag_uuid} removed from {removed_count} watches")
             except Exception as e:
@@ -98,6 +100,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
                 for watch_uuid, watch in datastore.data['watching'].items():
                     if watch.get('tags') and tag_uuid in watch['tags']:
                         watch['tags'].remove(tag_uuid)
+                        watch.commit()
                         unlinked_count += 1
                 logger.info(f"Background: Tag {tag_uuid} unlinked from {unlinked_count} watches")
             except Exception as e:
@@ -114,6 +117,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
     def delete_all():
         # Clear all tags from settings immediately
         datastore.data['settings']['application']['tags'] = {}
+        datastore.commit()
 
         # Clear tags from all watches in background thread to avoid blocking
         def clear_all_tags_background():
@@ -122,6 +126,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
             try:
                 for watch_uuid, watch in datastore.data['watching'].items():
                     watch['tags'] = []
+                    watch.commit()
                     cleared_count += 1
                 logger.info(f"Background: Cleared tags from {cleared_count} watches")
             except Exception as e:
@@ -216,7 +221,7 @@ def construct_blueprint(datastore: ChangeDetectionStore):
 
         datastore.data['settings']['application']['tags'][uuid].update(form.data)
         datastore.data['settings']['application']['tags'][uuid]['processor'] = 'restock_diff'
-        datastore.needs_write_urgent = True
+        datastore.commit()
         flash(gettext("Updated"))
 
         return redirect(url_for('tags.tags_overview_page'))

changedetectionio/blueprint/ui/__init__.py

@@ -24,7 +24,7 @@ def _handle_operations(op, uuids, datastore, worker_pool, update_q, queuedWatchM
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid]['paused'] = True
-                datastore.mark_watch_dirty(uuid)
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches paused").format(len(uuids)))
 
@@ -32,7 +32,7 @@ def _handle_operations(op, uuids, datastore, worker_pool, update_q, queuedWatchM
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid.strip()]['paused'] = False
-                datastore.mark_watch_dirty(uuid)
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches unpaused").format(len(uuids)))
 
@@ -47,7 +47,7 @@ def _handle_operations(op, uuids, datastore, worker_pool, update_q, queuedWatchM
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid]['notification_muted'] = True
-                datastore.mark_watch_dirty(uuid)
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches muted").format(len(uuids)))
 
@@ -55,7 +55,7 @@ def _handle_operations(op, uuids, datastore, worker_pool, update_q, queuedWatchM
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid]['notification_muted'] = False
-                datastore.mark_watch_dirty(uuid)
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches un-muted").format(len(uuids)))
 
@@ -71,7 +71,7 @@ def _handle_operations(op, uuids, datastore, worker_pool, update_q, queuedWatchM
         for uuid in uuids:
             if datastore.data['watching'].get(uuid):
                 datastore.data['watching'][uuid]["last_error"] = False
-                datastore.mark_watch_dirty(uuid)
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches errors cleared").format(len(uuids)))
 
@@ -92,6 +92,7 @@ def _handle_operations(op, uuids, datastore, worker_pool, update_q, queuedWatchM
                 datastore.data['watching'][uuid]['notification_body'] = None
                 datastore.data['watching'][uuid]['notification_urls'] = []
                 datastore.data['watching'][uuid]['notification_format'] = USE_SYSTEM_DEFAULT_NOTIFICATION_FORMAT_FOR_WATCH
+                datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches set to use default notification settings").format(len(uuids)))
 
@@ -107,6 +108,7 @@ def _handle_operations(op, uuids, datastore, worker_pool, update_q, queuedWatchM
                             datastore.data['watching'][uuid]['tags'] = []
 
                         datastore.data['watching'][uuid]['tags'].append(tag_uuid)
+                        datastore.data['watching'][uuid].commit()
         if emit_flash:
             flash(gettext("{} watches were tagged").format(len(uuids)))
 

changedetectionio/blueprint/ui/diff.py

@@ -100,23 +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
@@ -156,23 +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
@@ -212,24 +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
@@ -279,38 +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

@@ -71,8 +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])
 
@@ -149,58 +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
@@ -240,7 +197,11 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
 
             # Recast it if need be to right data Watch handler
             watch_class = processors.get_custom_watch_obj_for_processor(form.data.get('processor'))
-            datastore.data['watching'][uuid] = watch_class(datastore_path=datastore.datastore_path, default=datastore.data['watching'][uuid])
+            datastore.data['watching'][uuid] = watch_class(datastore_path=datastore.datastore_path, __datastore=datastore.data, default=datastore.data['watching'][uuid])
+
+            # Save the watch immediately
+            datastore.data['watching'][uuid].commit()
+
             flash(gettext("Updated watch - unpaused!") if request.args.get('unpause_on_save') else gettext("Updated watch."))
 
             # Cleanup any browsersteps session for this watch
@@ -250,10 +211,6 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
             except Exception as e:
                 logger.debug(f"Error cleaning up browsersteps session: {e}")
 
-            # Re #286 - We wait for syncing new data to disk in another thread every 60 seconds
-            # But in the case something is added we should save straight away
-            datastore.needs_write_urgent = True
-
             # Do not queue on edit if its not within the time range
 
             # @todo maybe it should never queue anyway on edit...
@@ -310,6 +267,13 @@ 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')}"]
@@ -422,6 +386,9 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
                     s = re.sub(r'[0-9]+', r'\\d+', s)
                     datastore.data["watching"][uuid]['ignore_text'].append('/' + s + '/')
 
+            # Save the updated ignore_text
+            datastore.data["watching"][uuid].commit()
+
         return f"<a href={url_for('ui.ui_preview.preview_page', uuid=uuid)}>Click to preview</a>"
     
     return edit_blueprint

changedetectionio/blueprint/ui/preview.py

@@ -38,24 +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 = []
@@ -160,39 +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

@@ -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:
@@ -38,4 +39,4 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
 
         return redirect(url_for('watchlist.index', tag=request.args.get('tag','')))
 
-    return views_blueprint
+    return views_blueprint

changedetectionio/blueprint/watchlist/__init__.py

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

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

@@ -1,5 +1,9 @@
 {%- extends 'base.html' -%}
 {%- block content -%}
+{%- set tips = [
+    _("Changedetection.io can monitor more than just web-pages! See our plugins!") ~ ' <a href="https://changedetection.io/plugins">' ~ _('More info') ~ '</a>',
+    _("You can also add 'shared' watches.") ~ ' <a href="https://github.com/dgtlmoon/changedetection.io/wiki/Sharing-a-Watch">' ~ _('More info') ~ '</a>'
+] -%}
 {%- from '_helpers.html' import render_simple_field, render_field, render_nolabel_field, sort_by_title -%}
 <script src="{{url_for('static_content', group='js', filename='jquery-3.6.0.min.js')}}"></script>
 <script src="{{url_for('static_content', group='js', filename='watch-overview.js')}}" defer></script>
@@ -10,6 +14,46 @@
 // Initialize Feather icons after the page loads
 document.addEventListener('DOMContentLoaded', function() {
     feather.replace();
+
+    // Intersection Observer for lazy loading favicons
+    // Only load favicon images when they enter the viewport
+    if ('IntersectionObserver' in window) {
+        const faviconObserver = new IntersectionObserver((entries, observer) => {
+            entries.forEach(entry => {
+                if (entry.isIntersecting) {
+                    const img = entry.target;
+                    const src = img.getAttribute('data-src');
+
+                    if (src) {
+                        // Load the actual favicon
+                        img.src = src;
+                        img.removeAttribute('data-src');
+                    }
+
+                    // Stop observing this image
+                    observer.unobserve(img);
+                }
+            });
+        }, {
+            // Start loading slightly before the image enters viewport
+            rootMargin: '50px',
+            threshold: 0.01
+        });
+
+        // Observe all lazy favicon images
+        document.querySelectorAll('.lazy-favicon').forEach(img => {
+            faviconObserver.observe(img);
+        });
+    } else {
+        // Fallback for older browsers: load all favicons immediately
+        document.querySelectorAll('.lazy-favicon').forEach(img => {
+            const src = img.getAttribute('data-src');
+            if (src) {
+                img.src = src;
+                img.removeAttribute('data-src');
+            }
+        });
+    }
 });
 </script>
 <style>
@@ -69,7 +113,9 @@ html[data-darkmode="true"] .watch-tag-list.tag-{{ class_name }} {
             </div>
 
         </fieldset>
-        <span style="color:#eee; font-size: 80%;"><img alt="{{ _('Create a shareable link') }}" style="height: 1em;display:inline-block;" src="{{url_for('static_content', group='images', filename='spread-white.svg')}}" > {{ _("Tip: You can also add 'shared' watches.") }} <a href="https://github.com/dgtlmoon/changedetection.io/wiki/Sharing-a-Watch">{{ _('More info') }}</a></span>
+        <span style="color:#eee; font-size: 80%;">
+            <strong>Tip: </strong> {{ tips | random | safe }}
+        </span>
     </form>
 </div>
 <div class="box">
@@ -200,28 +246,38 @@ html[data-darkmode="true"] .watch-tag-list.tag-{{ class_name }} {
                 <td class="title-col inline">
                     <div class="flex-wrapper">
                     {% if 'favicons_enabled' not in ui_settings or ui_settings['favicons_enabled'] %}
-                        <div>{# A page might have hundreds of these images, set IMG options for lazy loading, don't set SRC if we dont have it so it doesnt fetch the placeholder'  #}
-                            <img alt="Favicon thumbnail" class="favicon" loading="lazy" decoding="async" fetchpriority="low" {% if favicon %} src="{{url_for('static_content', group='favicon', filename=watch.uuid)}}" {% else %} src='data:image/svg+xml;utf8,%3Csvg xmlns="http://www.w3.org/2000/svg" width="7.087" height="7.087" viewBox="0 0 7.087 7.087"%3E%3Ccircle cx="3.543" cy="3.543" r="3.279" stroke="%23e1e1e1" stroke-width="0.45" fill="none" opacity="0.74"/%3E%3C/svg%3E' {%  endif %} >
+                        <div>
+                            {# Intersection Observer lazy loading: store real URL in data-src, load only when visible in viewport #}
+                            <img alt="Favicon thumbnail"
+                                 class="favicon lazy-favicon"
+                                 loading="lazy"
+                                 decoding="async"
+                                 fetchpriority="low"
+                                 {% if favicon %}
+                                 data-src="{{url_for('static_content', group='favicon', filename=watch.uuid)}}"
+                                 {% endif %}
+                                 src='data:image/svg+xml;utf8,%3Csvg xmlns="http://www.w3.org/2000/svg" width="7.087" height="7.087" viewBox="0 0 7.087 7.087"%3E%3Ccircle cx="3.543" cy="3.543" r="3.279" stroke="%23e1e1e1" stroke-width="0.45" fill="none" opacity="0.74"/%3E%3C/svg%3E'>
                         </div>
                     {%  endif %}
                         <div>
-                        <span class="watch-title">
-                            {% if system_use_url_watchlist or watch.get('use_page_title_in_list') %}
-                                {{ watch.label }}
-                            {% else %}
-                                {{ watch.get('title') or watch.link }}
-                            {% endif %}
-                           <a class="external" target="_blank" rel="noopener" href="{{ watch.link.replace('source:','') }}">&nbsp;</a>
-                        </span>
+                            {%- if watch['processor'] and watch['processor'] in processor_badge_texts -%}
+                                <span class="processor-badge processor-badge-{{ watch['processor'] }}" title="{{ processor_descriptions.get(watch['processor'], watch['processor']) }}">{{ processor_badge_texts[watch['processor']] }}</span>
+                            {%- endif -%}
+                            <span class="watch-title">
+                                {% if system_use_url_watchlist or watch.get('use_page_title_in_list') %}
+                                    {{ watch.label }}
+                                {% else %}
+                                    {{ watch.get('title') or watch.link }}
+                                {% endif %}
+                               <a class="external" target="_blank" rel="noopener" href="{{ watch.link.replace('source:','') }}">&nbsp;</a>
+                            </span>
                             <div class="error-text" style="display:none;">{{ watch.compile_error_texts(has_proxies=datastore.proxy_list)|safe }}</div>
                             {%- if watch['processor'] == 'text_json_diff'  -%}
                                 {%- if watch['has_ldjson_price_data'] and not watch['track_ldjson_price_data']  -%}
                                 <div class="ldjson-price-track-offer">Switch to Restock & Price watch mode? <a href="{{url_for('price_data_follower.accept', uuid=watch.uuid)}}" class="pure-button button-xsmall">Yes</a> <a href="{{url_for('price_data_follower.reject', uuid=watch.uuid)}}" class="">No</a></div>
                                 {%- endif -%}
                             {%- endif -%}
-                            {%- if watch['processor'] and watch['processor'] in processor_badge_texts -%}
-                                <span class="processor-badge processor-badge-{{ watch['processor'] }}" title="{{ processor_descriptions.get(watch['processor'], watch['processor']) }}">{{ processor_badge_texts[watch['processor']] }}</span>
-                            {%- endif -%}
+
                             {%- for watch_tag_uuid, watch_tag in datastore.get_all_tags_for_watch(watch['uuid']).items() -%}
                               <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 -%}

changedetectionio/content_fetchers/playwright.py

@@ -1,3 +1,4 @@
+import asyncio
 import gc
 import json
 import os
@@ -349,12 +350,7 @@ class fetcher(Fetcher):
 
             if self.status_code != 200 and not ignore_status_codes:
                 screenshot = await capture_full_page_async(self.page, screenshot_format=self.screenshot_format, watch_uuid=watch_uuid, lock_viewport_elements=self.lock_viewport_elements)
-                # 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()
+                # 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:
@@ -370,12 +366,7 @@ class fetcher(Fetcher):
                     try:
                         await self.iterate_browser_steps(start_url=url)
                     except BrowserStepsStepException:
-                        try:
-                            await context.close()
-                            await browser.close()
-                        except Exception as e:
-                            # Fine, could be messy situation
-                            pass
+                        # Finally block will handle cleanup
                         raise
 
                     await self.page.wait_for_timeout(extra_wait * 1000)
@@ -424,35 +415,40 @@ class fetcher(Fetcher):
                 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
@@ -221,19 +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,
@@ -263,9 +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:
@@ -274,7 +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}"))
@@ -343,6 +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!")
@@ -368,7 +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
@@ -437,15 +476,9 @@ class fetcher(Fetcher):
         logger.debug(f"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 aggressive memory cleanup - pyppeteer base64 decode creates temporary buffers
+        # Force garbage collection - pyppeteer base64 decode creates temporary buffers
         import gc
         gc.collect()
-        # Release C-level memory from base64 decode back to OS
-        try:
-            import ctypes
-            ctypes.CDLL('libc.so.6').malloc_trim(0)
-        except Exception:
-            pass
         self.xpath_data = await self.page.evaluate(XPATH_ELEMENT_JS, {
             "visualselector_xpath_selectors": visualselector_xpath_selectors,
             "max_height": MAX_TOTAL_HEIGHT

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

@@ -46,6 +46,7 @@ class model(dict):
                     'filter_failure_notification_threshold_attempts': _FILTER_FAILURE_THRESHOLD_ATTEMPTS_DEFAULT,
                     'global_ignore_text': [], # List of text to ignore when calculating the comparison checksum
                     'global_subtractive_selectors': [],
+                    'history_snapshot_max_length': None,
                     'ignore_whitespace': True,
                     'ignore_status_codes': False, #@todo implement, as ternary.
                     'ssim_threshold': '0.96',  # Default SSIM threshold for screenshot comparison

changedetectionio/model/Tag.py

@@ -1,10 +1,52 @@
+"""
+Tag/Group domain model for organizing and overriding watch settings.
+
+ARCHITECTURE NOTE: Configuration Override Hierarchy
+===================================================
+
+Tags can override Watch settings when overrides_watch=True.
+Current implementation requires manual checking in processors:
+
+    for tag_uuid in watch.get('tags'):
+        tag = datastore['settings']['application']['tags'][tag_uuid]
+        if tag.get('overrides_watch'):
+            restock_settings = tag.get('restock_settings', {})
+            break
+
+With Pydantic, this would be automatic via chain resolution:
+    Watch → Tag (first with overrides_watch) → Global
+
+See: Watch.py model docstring for full Pydantic architecture explanation
+See: processors/restock_diff/processor.py:184-192 for current manual implementation
+"""
 
 from changedetectionio.model import watch_base
 
 
 class model(watch_base):
+    """
+    Tag domain model - groups watches and can override their settings.
+
+    Tags inherit from watch_base to reuse all the same fields as Watch.
+    When overrides_watch=True, tag settings take precedence over watch settings
+    for all watches in this tag/group.
+
+    Fields:
+        overrides_watch (bool): If True, this tag's settings override watch settings
+        title (str): Display name for this tag/group
+        uuid (str): Unique identifier
+        ... (all fields from watch_base can be set as tag-level overrides)
+
+    Resolution order when overrides_watch=True:
+        Watch.field → Tag.field (if overrides_watch) → Global.field
+    """
 
     def __init__(self, *arg, **kw):
+        # Store datastore reference (optional for Tags, but good for consistency)
+        self.__datastore = kw.get('__datastore')
+        if kw.get('__datastore'):
+            del kw['__datastore']
+
         super(model, self).__init__(*arg, **kw)
 
         self['overrides_watch'] = kw.get('default', {}).get('overrides_watch')

changedetectionio/model/Watch.py

@@ -1,3 +1,32 @@
+"""
+Watch domain model for change detection monitoring.
+
+ARCHITECTURE NOTE: Configuration Override Hierarchy
+===================================================
+
+This module implements Watch objects that inherit from dict (technical debt).
+The dream architecture would use Pydantic for:
+
+1. CHAIN RESOLUTION (Watch → Tag → Global Settings)
+   - Current: Manual resolution scattered across codebase
+   - Future: @computed_field properties with automatic resolution
+   - Examples: resolved_fetch_backend, resolved_restock_settings, etc.
+
+2. DATABASE BACKEND ABSTRACTION
+   - Current: Domain model tightly coupled to file-based JSON storage
+   - Future: Domain model (Pydantic) separate from persistence layer
+   - Enables: Easy migration to PostgreSQL, MongoDB, etc.
+
+3. TYPE SAFETY & VALIDATION
+   - Current: Dict access with no compile-time checks
+   - Future: Type hints, IDE autocomplete, validation at boundaries
+
+See class model docstring for detailed explanation and examples.
+See: processors/restock_diff/processor.py:184-192 for manual resolution example
+"""
+import gc
+from copy import copy
+
 from blinker import signal
 from changedetectionio.validate_url import is_safe_valid_url
 
@@ -13,7 +42,7 @@ 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}
@@ -101,6 +130,99 @@ def _brotli_save(contents, filepath, mode=None, fallback_uncompressed=False):
 
 
 class model(watch_base):
+    """
+    Watch domain model for monitoring URL changes.
+
+    Inherits from watch_base (which inherits dict) - see watch_base docstring for field documentation.
+
+    ## Configuration Override Hierarchy (Chain Resolution)
+
+    The dream architecture uses a 3-level resolution chain:
+        Watch settings → Tag/Group settings → Global settings
+
+    Current implementation is MANUAL (see processor.py:184-192 for example):
+        - Processors manually check watch.get('field')
+        - Then loop through watch.tags to find first tag with overrides_watch=True
+        - Finally fall back to datastore['settings']['application']['field']
+
+    FUTURE: Pydantic-based chain resolution would enable:
+
+        ```python
+        # Instead of manual resolution in every processor:
+        restock_settings = watch.get('restock_settings', {})
+        for tag_uuid in watch.get('tags'):
+            tag = datastore['settings']['application']['tags'][tag_uuid]
+            if tag.get('overrides_watch'):
+                restock_settings = tag.get('restock_settings', {})
+                break
+
+        # Clean computed properties with automatic resolution:
+        @computed_field
+        def resolved_restock_settings(self) -> dict:
+            if self.restock_settings:
+                return self.restock_settings
+            for tag_uuid in self.tags:
+                tag = self._datastore.get_tag(tag_uuid)
+                if tag.overrides_watch and tag.restock_settings:
+                    return tag.restock_settings
+            return self._datastore.settings.restock_settings or {}
+
+        # Usage: watch.resolved_restock_settings (automatic, type-safe, tested once)
+        ```
+
+    Benefits of Pydantic migration:
+        1. Single source of truth for resolution logic (not scattered across processors)
+        2. Type safety + IDE autocomplete (watch.resolved_fetch_backend vs dict navigation)
+        3. Database backend abstraction (domain model separate from persistence)
+        4. Automatic validation at boundaries
+        5. Self-documenting via type hints
+        6. Easy to test resolution independently
+
+    Resolution chain examples that would benefit:
+        - fetch_backend: watch → tag → global (see get_fetch_backend property)
+        - notification_urls: watch → tag → global
+        - time_between_check: watch → global (see threshold_seconds)
+        - restock_settings: watch → tag (see processors/restock_diff/processor.py:184-192)
+        - history_snapshot_max_length: watch → global (see save_history_blob:550-556)
+        - All processor_config_* settings could use tag overrides
+
+    ## Database Backend Abstraction with Pydantic
+
+    Current: Watch inherits dict, tightly coupled to file-based JSON storage
+    Future: Domain model (Watch) separate from persistence layer
+
+        ```python
+        # Domain model (database-agnostic)
+        class Watch(BaseModel):
+            uuid: str
+            url: str
+            # ... validation, business logic
+
+        # Pluggable backends
+        class DataStoreBackend(ABC):
+            def save_watch(self, watch: Watch): ...
+            def load_watch(self, uuid: str) -> Watch: ...
+
+        # Implementations: FileBackend, MongoBackend, PostgresBackend, etc.
+        ```
+
+    This would enable:
+        - Easy migration between storage backends (file → postgres → mongodb)
+        - Pydantic handles serialization/deserialization automatically
+        - Domain logic stays clean (no storage concerns in Watch methods)
+
+    ## Migration Path
+
+    Given existing codebase, incremental migration recommended:
+        1. Create Pydantic models alongside existing dict-based models
+        2. Add .to_pydantic() / .from_pydantic() bridge methods
+        3. Gradually migrate code to use Pydantic models
+        4. Remove dict inheritance once migration complete
+
+    See: watch_base docstring for technical debt discussion
+    See: processors/restock_diff/processor.py:184-192 for manual resolution example
+    See: Watch.py:550-556 for nested dict navigation that would become watch.resolved_*
+    """
     __newest_history_key = None
     __history_n = 0
     jitter_seconds = 0
@@ -109,8 +231,15 @@ 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']
@@ -121,6 +250,9 @@ class model(watch_base):
         # Be sure the cached timestamp is ready
         bump = self.history
 
+    # Note: __deepcopy__, __getstate__, and __setstate__ are inherited from watch_base
+    # This prevents memory leaks by sharing __datastore reference instead of copying it
+
     @property
     def viewed(self):
         # Don't return viewed when last_viewed is 0 and newest_key is 0
@@ -230,8 +362,30 @@ class model(watch_base):
     @property
     def get_fetch_backend(self):
         """
-        Like just using the `fetch_backend` key but there could be some logic
-        :return:
+        Get the fetch backend for this watch with special case handling.
+
+        CHAIN RESOLUTION OPPORTUNITY:
+        Currently returns watch.fetch_backend directly, but doesn't implement
+        Watch → Tag → Global resolution chain. With Pydantic:
+
+        @computed_field
+        def resolved_fetch_backend(self) -> str:
+            # Special case: PDFs always use html_requests
+            if self.is_pdf:
+                return 'html_requests'
+
+            # Watch override
+            if self.fetch_backend and self.fetch_backend != 'system':
+                return self.fetch_backend
+
+            # Tag override (first tag with overrides_watch=True wins)
+            for tag_uuid in self.tags:
+                tag = self._datastore.get_tag(tag_uuid)
+                if tag.overrides_watch and tag.fetch_backend:
+                    return tag.fetch_backend
+
+            # Global default
+            return self._datastore.settings.fetch_backend
         """
         # Maybe also if is_image etc?
         # This is because chrome/playwright wont render the PDF in the browser and we will just fetch it and use pdf2html to see the text.
@@ -423,16 +577,49 @@ 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()
@@ -441,7 +628,6 @@ 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
@@ -501,6 +687,20 @@ class model(watch_base):
         self.__newest_history_key = timestamp
         self.__history_n += 1
 
+        # MANUAL CHAIN RESOLUTION: Watch → Global
+        # With Pydantic, this would become: maxlen = watch.resolved_history_snapshot_max_length
+        # @computed_field def resolved_history_snapshot_max_length(self) -> Optional[int]:
+        #     if self.history_snapshot_max_length: return self.history_snapshot_max_length
+        #     if tag := self._get_override_tag(): return tag.history_snapshot_max_length
+        #     return self._datastore.settings.history_snapshot_max_length
+        maxlen = (
+                self.get('history_snapshot_max_length')
+                or (self.__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
 
@@ -613,6 +813,11 @@ 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:
@@ -629,20 +834,32 @@ 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.
@@ -773,6 +990,57 @@ class model(watch_base):
     def toggle_mute(self):
         self['notification_muted'] ^= True
 
+    def commit(self):
+        """
+        Save this watch immediately to disk using atomic write.
+
+        Replaces the old dirty-tracking system with immediate persistence.
+        Uses atomic write pattern (temp file + rename) for crash safety.
+
+        Fire-and-forget: Logs errors but does not raise exceptions.
+        Watch data remains in memory even if save fails, so next commit will retry.
+        """
+        from loguru import logger
+
+        if not self.__datastore:
+            logger.error(f"Cannot commit watch {self.get('uuid')} without datastore reference")
+            return
+
+        if not self.watch_data_dir:
+            logger.error(f"Cannot commit watch {self.get('uuid')} without datastore_path")
+            return
+
+        # Convert to dict for serialization, excluding processor config keys
+        # Processor configs are stored separately in processor-specific JSON files
+        # Use deepcopy to prevent mutations from affecting the original Watch object
+        import copy
+
+        # Acquire datastore lock to prevent concurrent modifications during copy
+        # Take a quick shallow snapshot under lock, then deep copy outside lock
+        lock = self.__datastore.lock if self.__datastore and hasattr(self.__datastore, 'lock') else None
+
+        if lock:
+            with lock:
+                snapshot = dict(self)
+        else:
+            snapshot = dict(self)
+
+        # Deep copy snapshot (slower, but done outside lock to minimize contention)
+        watch_dict = {k: copy.deepcopy(v) for k, v in snapshot.items() if not k.startswith('processor_config_')}
+
+        # Normalize browser_steps: if no meaningful steps, save as empty list
+        if not self.has_browser_steps:
+            watch_dict['browser_steps'] = []
+
+        # Use existing atomic write helper
+        from changedetectionio.store.file_saving_datastore import save_watch_atomic
+        try:
+            save_watch_atomic(self.watch_data_dir, self.get('uuid'), watch_dict)
+            logger.debug(f"Committed watch {self.get('uuid')}")
+        except Exception as e:
+            logger.error(f"Failed to commit watch {self.get('uuid')}: {e}")
+
+
     def extra_notification_token_values(self):
         # Used for providing extra tokens
         # return {'widget': 555}

changedetectionio/model/__init__.py

@@ -6,6 +6,147 @@ USE_SYSTEM_DEFAULT_NOTIFICATION_FORMAT_FOR_WATCH = 'System default'
 CONDITIONS_MATCH_LOGIC_DEFAULT = 'ALL'
 
 class watch_base(dict):
+    """
+    Base watch domain model (inherits from dict for backward compatibility).
+
+    WARNING: This class inherits from dict, which violates proper encapsulation.
+    Dict inheritance is legacy technical debt that should be refactored to a proper
+    domain model (e.g., Pydantic BaseModel) for better type safety and validation.
+
+    TODO: Migrate to Pydantic BaseModel for:
+          - Type safety and IDE autocomplete
+          - Automatic validation
+          - Clear separation between domain model and serialization
+          - Database backend abstraction (file → postgres → mongodb)
+          - Configuration override chain resolution (Watch → Tag → Global)
+          - Immutability options
+          - Better testing
+
+    CHAIN RESOLUTION ARCHITECTURE:
+        The dream is a 3-level override hierarchy:
+            Watch settings → Tag/Group settings → Global settings
+
+        Current implementation: MANUAL resolution scattered across codebase
+        - Processors manually check watch.get('field')
+        - Loop through tags to find overrides_watch=True
+        - Fall back to datastore['settings']['application']['field']
+
+        Pydantic implementation: AUTOMATIC resolution via @computed_field
+        - Single source of truth for each setting's resolution logic
+        - Type-safe, testable, self-documenting
+        - Example: watch.resolved_fetch_backend (instead of nested dict navigation)
+
+        See: Watch.py model docstring for detailed Pydantic architecture plan
+        See: Tag.py model docstring for tag override explanation
+        See: processors/restock_diff/processor.py:184-192 for current manual example
+
+    Core Fields:
+        uuid (str): Unique identifier for this watch (auto-generated)
+        url (str): Target URL to monitor for changes
+        title (str|None): Custom display name (overrides page_title if set)
+        page_title (str|None): Title extracted from <title> tag of monitored page
+        tags (List[str]): List of tag UUIDs for categorization
+        tag (str): DEPRECATED - Old single-tag system, use tags instead
+
+    Check Configuration:
+        processor (str): Processor type ('text_json_diff', 'restock_diff', etc.)
+        fetch_backend (str): Fetcher to use ('system', 'html_requests', 'playwright', etc.)
+        method (str): HTTP method ('GET', 'POST', etc.)
+        headers (dict): Custom HTTP headers to send
+        proxy (str|None): Preferred proxy server
+        paused (bool): Whether change detection is paused
+
+    Scheduling:
+        time_between_check (dict): Check interval {'weeks': int, 'days': int, 'hours': int, 'minutes': int, 'seconds': int}
+        time_between_check_use_default (bool): Use global default interval if True
+        time_schedule_limit (dict): Weekly schedule limiting when checks can run
+            Structure: {
+                'enabled': bool,
+                'monday/tuesday/.../sunday': {
+                    'enabled': bool,
+                    'start_time': str ('HH:MM'),
+                    'duration': {'hours': str, 'minutes': str}
+                }
+            }
+
+    Content Filtering:
+        include_filters (List[str]): CSS/XPath selectors to extract content
+        subtractive_selectors (List[str]): Selectors to remove from content
+        ignore_text (List[str]): Text patterns to ignore in change detection
+        trigger_text (List[str]): Text/regex that must be present to trigger change
+        text_should_not_be_present (List[str]): Text that should NOT be present
+        extract_text (List[str]): Regex patterns to extract specific text after filtering
+
+    Text Processing:
+        trim_text_whitespace (bool): Strip leading/trailing whitespace
+        sort_text_alphabetically (bool): Sort lines alphabetically before comparison
+        remove_duplicate_lines (bool): Remove duplicate lines
+        check_unique_lines (bool): Compare against all history for unique lines
+        strip_ignored_lines (bool|None): Remove lines matching ignore patterns
+
+    Change Detection Filters:
+        filter_text_added (bool): Include added text in change detection
+        filter_text_removed (bool): Include removed text in change detection
+        filter_text_replaced (bool): Include replaced text in change detection
+
+    Browser Automation:
+        browser_steps (List[dict]): Browser automation steps for JS-heavy sites
+        browser_steps_last_error_step (int|None): Last step that caused error
+        webdriver_delay (int|None): Seconds to wait after page load
+        webdriver_js_execute_code (str|None): JavaScript to execute before extraction
+
+    Restock Detection:
+        in_stock_only (bool): Only trigger on in-stock transitions
+        follow_price_changes (bool): Monitor price changes
+        has_ldjson_price_data (bool|None): Whether page has LD-JSON price data
+        track_ldjson_price_data (str|None): Track LD-JSON price data ('ACCEPT', 'REJECT', None)
+        price_change_threshold_percent (float|None): Minimum price change % to trigger
+
+    Notifications:
+        notification_urls (List[str]): Apprise URLs for notifications
+        notification_title (str|None): Custom notification title template
+        notification_body (str|None): Custom notification body template
+        notification_format (str): Notification format (e.g., 'System default', 'Text', 'HTML')
+        notification_muted (bool): Disable notifications for this watch
+        notification_screenshot (bool): Include screenshot in notifications
+        notification_alert_count (int): Number of notifications sent
+        last_notification_error (str|None): Last notification error message
+        body (str|None): DEPRECATED? Legacy notification body field
+        filter_failure_notification_send (bool): Send notification on filter failures
+
+    History & State:
+        date_created (int|None): Unix timestamp of watch creation
+        last_checked (int): Unix timestamp of last check
+        last_viewed (int): History snapshot key of last user view
+        last_error (str|bool): Last error message or False if no error
+        check_count (int): Total number of checks performed
+        fetch_time (float): Duration of last fetch in seconds
+        consecutive_filter_failures (int): Counter for consecutive filter match failures
+        previous_md5 (str|bool): MD5 hash of previous content
+        previous_md5_before_filters (str|bool): MD5 hash before filters applied
+        history_snapshot_max_length (int|None): Max history snapshots to keep (None = use global)
+
+    Conditions:
+        conditions (dict): Custom conditions for change detection logic
+        conditions_match_logic (str): Logic operator ('ALL', 'ANY') for conditions
+
+    Metadata:
+        content-type (str|None): Content-Type from last fetch
+        remote_server_reply (str|None): Server header from last response
+        ignore_status_codes (List[int]|None): HTTP status codes to ignore
+        use_page_title_in_list (bool|None): Display page title in watch list (None = use system default)
+
+    Instance Attributes (not serialized):
+        __datastore: Reference to parent DataStore (set externally after creation)
+        watch_data_dir: Filesystem path for this watch's data directory
+
+    Notes:
+        - Many fields default to None to distinguish "not set" from "set to default"
+        - When field is None, system-level defaults are used
+        - Processor-specific configs (e.g., processor_config_*) are NOT stored in watch.json
+          They are stored in separate {processor_name}.json files
+        - This class is used for both Watch and Tag objects (tags reuse the structure)
+    """
 
     def __init__(self, *arg, **kw):
         self.update({
@@ -32,6 +173,7 @@ class watch_base(dict):
             'filter_text_replaced': True,
             'follow_price_changes': True,
             'has_ldjson_price_data': None,
+            'history_snapshot_max_length': None,
             'headers': {},  # Extra headers to send
             'ignore_text': [],  # List of text to ignore when calculating the comparison checksum
             'ignore_status_codes': None,
@@ -139,4 +281,100 @@ class watch_base(dict):
         super(watch_base, self).__init__(*arg, **kw)
 
         if self.get('default'):
-            del self['default']
+            del self['default']
+
+    def __deepcopy__(self, memo):
+        """
+        Custom deepcopy for all watch_base subclasses (Watch, Tag, etc.).
+
+        CRITICAL FIX: Prevents copying large reference objects like __datastore
+        which would cause exponential memory growth when Watch objects are deepcopied.
+
+        This is called by:
+        - api/Watch.py:76 (API endpoint)
+        - api/Tags.py:28 (Tags API)
+        - processors/base.py:26 (EVERY processor run)
+        - store/__init__.py:544 (clone watch)
+        - And other locations
+        """
+        from copy import deepcopy
+
+        # Create new instance without calling __init__
+        cls = self.__class__
+        new_obj = cls.__new__(cls)
+        memo[id(self)] = new_obj
+
+        # Copy the dict data (all the settings)
+        for key, value in self.items():
+            new_obj[key] = deepcopy(value, memo)
+
+        # Copy instance attributes dynamically
+        # This handles Watch-specific attrs (like __datastore) and any future subclass attrs
+        for attr_name in dir(self):
+            # Skip methods, special attrs, and dict keys
+            if attr_name.startswith('_') and not attr_name.startswith('__'):
+                # This catches _model__datastore, _model__history_n, etc.
+                try:
+                    attr_value = getattr(self, attr_name)
+
+                    # Special handling: Share references to large objects instead of copying
+                    # Examples: __datastore, __app_reference, __global_settings, etc.
+                    if attr_name.endswith('__datastore') or attr_name.endswith('__app'):
+                        # Share the reference (don't copy!) to prevent memory leaks
+                        setattr(new_obj, attr_name, attr_value)
+                    # Skip cache attributes - let them regenerate on demand
+                    elif 'cache' in attr_name.lower():
+                        pass  # Don't copy caches
+                    # Copy regular instance attributes
+                    elif not callable(attr_value):
+                        setattr(new_obj, attr_name, attr_value)
+                except AttributeError:
+                    pass  # Attribute doesn't exist in this instance
+
+        return new_obj
+
+    def __getstate__(self):
+        """
+        Custom pickle serialization for all watch_base subclasses.
+
+        Excludes large reference objects (like __datastore) from serialization.
+        """
+        # Get the dict data
+        state = dict(self)
+
+        # Collect instance attributes (excluding methods and large references)
+        instance_attrs = {}
+        for attr_name in dir(self):
+            if attr_name.startswith('_') and not attr_name.startswith('__'):
+                try:
+                    attr_value = getattr(self, attr_name)
+                    # Exclude large reference objects and caches from serialization
+                    if not (attr_name.endswith('__datastore') or
+                           attr_name.endswith('__app') or
+                           'cache' in attr_name.lower() or
+                           callable(attr_value)):
+                        instance_attrs[attr_name] = attr_value
+                except AttributeError:
+                    pass
+
+        if instance_attrs:
+            state['__instance_metadata__'] = instance_attrs
+
+        return state
+
+    def __setstate__(self, state):
+        """
+        Custom pickle deserialization for all watch_base subclasses.
+
+        WARNING: Large reference objects (like __datastore) are NOT restored!
+        Caller must restore these references after unpickling if needed.
+        """
+        # Extract metadata
+        metadata = state.pop('__instance_metadata__', {})
+
+        # Restore dict data
+        self.update(state)
+
+        # Restore instance attributes
+        for attr_name, attr_value in metadata.items():
+            setattr(self, attr_name, attr_value)

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/restock_diff/processor.py

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

changedetectionio/processors/text_json_diff/__init__.py

@@ -1,6 +1,12 @@
-
 from loguru import logger
 
+# Processor capabilities
+supports_visual_selector = True
+supports_browser_steps = True
+supports_text_filters_and_triggers = True
+supports_text_filters_and_triggers_elements = True
+supports_request_type = True
+
 
 
 def _task(watch, update_handler):
@@ -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/run_basic_tests.sh

@@ -67,7 +67,7 @@ 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
 FETCH_WORKERS=2 REMOVE_REQUESTS_OLD_SCREENSHOTS=false \
 pytest tests/test_*.py \
-  -n 18 \
+  -n 8 \
   --dist=load \
   -vvv \
   -s \

changedetectionio/store/__init__.py

@@ -9,18 +9,14 @@ from flask import (
 )
 from flask_babel import gettext
 
-from ..blueprint.rss import RSS_CONTENT_FORMAT_DEFAULT
-from ..html_tools import TRANSLATE_WHITESPACE_TABLE
-from ..model import App, Watch, USE_SYSTEM_DEFAULT_NOTIFICATION_FORMAT_FOR_WATCH
-from copy import deepcopy, copy
+from ..model import App, Watch
+from copy import deepcopy
 from os import path, unlink
-from threading import Lock
 import json
 import os
 import re
 import secrets
 import sys
-import threading
 import time
 import uuid as uuid_builder
 from loguru import logger
@@ -35,7 +31,6 @@ except ImportError:
     HAS_ORJSON = False
 
 from ..processors import get_custom_watch_obj_for_processor
-from ..processors.restock_diff import Restock
 
 # Import the base class and helpers
 from .file_saving_datastore import FileSavingDataStore, load_all_watches, save_watch_atomic, save_json_atomic
@@ -61,9 +56,7 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
         # Should only be active for docker
         # logging.basicConfig(filename='/dev/stdout', level=logging.INFO)
         self.datastore_path = datastore_path
-        self.needs_write = False
         self.start_time = time.time()
-        self.stop_thread = False
         self.save_version_copy_json_db(version_tag)
         self.reload_state(datastore_path=datastore_path, include_default_watches=include_default_watches, version_tag=version_tag)
 
@@ -137,6 +130,19 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
             )
             logger.info(f"Tag: {uuid} {tag['title']}")
 
+    def _rehydrate_watches(self):
+        """Rehydrate watch entities from stored data (converts dicts to Watch objects)."""
+        watch_count = len(self.__data.get('watching', {}))
+        if watch_count == 0:
+            return
+
+        logger.info(f"Rehydrating {watch_count} watches...")
+        watching_rehydrated = {}
+        for uuid, watch_dict in self.__data.get('watching', {}).items():
+            watching_rehydrated[uuid] = self.rehydrate_entity(uuid, watch_dict)
+        self.__data['watching'] = watching_rehydrated
+        logger.success(f"Rehydrated {watch_count} watches into Watch objects")
+
 
     def _load_state(self):
         """
@@ -174,7 +180,7 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
         self.json_store_path = os.path.join(self.datastore_path, "changedetection.json")
 
         # Base definition for all watchers (deepcopy part of #569)
-        self.generic_definition = deepcopy(Watch.model(datastore_path=datastore_path, default={}))
+        self.generic_definition = deepcopy(Watch.model(datastore_path=datastore_path, __datastore=self.__data, default={}))
 
         # Load build SHA if available (Docker deployments)
         if path.isfile('changedetectionio/source.txt'):
@@ -218,23 +224,29 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
                 logger.critical(f"Legacy datastore detected at {self.datastore_path}/url-watches.json")
                 logger.critical("Migration will be triggered via update_26")
 
-                # Load the legacy datastore to get its schema_version
+                # Load the legacy datastore
                 from .legacy_loader import load_legacy_format
                 legacy_path = os.path.join(self.datastore_path, "url-watches.json")
-                with open(legacy_path) as f:
-                    self.__data = json.load(f)
+                legacy_data = load_legacy_format(legacy_path)
 
-                if not self.__data:
+                if not legacy_data:
                     raise Exception("Failed to load legacy datastore from url-watches.json")
 
-                # update_26 will load the legacy data again and migrate to new format
-                # Only run updates AFTER the legacy schema version (e.g., if legacy is at 25, only run 26+)
+                # Store the loaded data
+                self.__data = legacy_data
+
+                # CRITICAL: Rehydrate watches from dicts into Watch objects
+                # This ensures watches have their methods available during migration
+                self._rehydrate_watches()
+
+                # update_26 will save watches to individual files and create changedetection.json
+                # Next startup will load from new format normally
                 self.run_updates()
 
 
             else:
                 # Fresh install - create new datastore
-                logger.critical(f"No datastore found, creating new datastore at {self.datastore_path}")
+                logger.warning(f"No datastore found, creating new datastore at {self.datastore_path}")
 
                 # Set schema version to latest (no updates needed)
                 updates_available = self.get_updates_available()
@@ -272,19 +284,19 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
                 self.__data['app_guid'] = "test-" + str(uuid_builder.uuid4())
             else:
                 self.__data['app_guid'] = str(uuid_builder.uuid4())
-            self.mark_settings_dirty()
+            self.commit()
 
         # Ensure RSS access token exists
         if not self.__data['settings']['application'].get('rss_access_token'):
             secret = secrets.token_hex(16)
             self.__data['settings']['application']['rss_access_token'] = secret
-            self.mark_settings_dirty()
+            self.commit()
 
         # Ensure API access token exists
         if not self.__data['settings']['application'].get('api_access_token'):
             secret = secrets.token_hex(16)
             self.__data['settings']['application']['api_access_token'] = secret
-            self.mark_settings_dirty()
+            self.commit()
 
         # Handle password reset lockfile
         password_reset_lockfile = os.path.join(self.datastore_path, "removepassword.lock")
@@ -292,9 +304,6 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
             self.remove_password()
             unlink(password_reset_lockfile)
 
-        # Start the background save thread
-        self.start_save_thread()
-
     def rehydrate_entity(self, uuid, entity, processor_override=None):
         """Set the dict back to the dict Watch object"""
         entity['uuid'] = uuid
@@ -308,7 +317,7 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
         if entity.get('processor') != 'text_json_diff':
             logger.trace(f"Loading Watch object '{watch_class.__module__}.{watch_class.__name__}' for UUID {uuid}")
 
-        entity = watch_class(datastore_path=self.datastore_path, default=entity)
+        entity = watch_class(datastore_path=self.datastore_path, __datastore=self.__data, default=entity)
         return entity
 
     # ============================================================================
@@ -361,22 +370,15 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
         Implementation of abstract method from FileSavingDataStore.
         Delegates to helper function and stores results in internal data structure.
         """
-        watching, watch_hashes = load_all_watches(
+        watching = load_all_watches(
             self.datastore_path,
-            self.rehydrate_entity,
-            self._compute_hash
+            self.rehydrate_entity
         )
 
         # Store loaded data
         self.__data['watching'] = watching
-        self._watch_hashes = watch_hashes
 
-        # Verify all watches have hashes
-        missing_hashes = [uuid for uuid in watching.keys() if uuid not in watch_hashes]
-        if missing_hashes:
-            logger.error(f"WARNING: {len(missing_hashes)} watches missing hashes after load: {missing_hashes[:5]}")
-        else:
-            logger.debug(f"All {len(watching)} watches have valid hashes")
+        logger.debug(f"Loaded {len(watching)} watches")
 
     def _delete_watch(self, uuid):
         """
@@ -400,7 +402,7 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
     def set_last_viewed(self, uuid, timestamp):
         logger.debug(f"Setting watch UUID: {uuid} last viewed to {int(timestamp)}")
         self.data['watching'][uuid].update({'last_viewed': int(timestamp)})
-        self.mark_watch_dirty(uuid)
+        self.data['watching'][uuid].commit()
 
         watch_check_update = signal('watch_check_update')
         if watch_check_update:
@@ -408,7 +410,22 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
 
     def remove_password(self):
         self.__data['settings']['application']['password'] = False
-        self.mark_settings_dirty()
+        self.commit()
+
+    def commit(self):
+        """
+        Save settings immediately to disk using atomic write.
+
+        Uses atomic write pattern (temp file + rename) for crash safety.
+
+        Fire-and-forget: Logs errors but does not raise exceptions.
+        Settings data remains in memory even if save fails, so next commit will retry.
+        """
+        try:
+            self._save_settings()
+            logger.debug("Committed settings")
+        except Exception as e:
+            logger.error(f"Failed to commit settings: {e}")
 
     def update_watch(self, uuid, update_obj):
 
@@ -427,7 +444,8 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
 
             self.__data['watching'][uuid].update(update_obj)
 
-        self.mark_watch_dirty(uuid)
+        # Immediate save
+        self.__data['watching'][uuid].commit()
 
     @property
     def threshold_seconds(self):
@@ -488,10 +506,6 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
                     except Exception as e:
                         logger.error(f"Failed to delete watch {watch_uuid} from storage: {e}")
 
-                    # Clean up tracking data
-                    self._watch_hashes.pop(watch_uuid, None)
-                    self._dirty_watches.discard(watch_uuid)
-
                     # Send delete signal
                     watch_delete_signal = signal('watch_deleted')
                     if watch_delete_signal:
@@ -513,21 +527,19 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
                 # Remove from watching dict
                 del self.data['watching'][uuid]
 
-                # Clean up tracking data
-                self._watch_hashes.pop(uuid, None)
-                self._dirty_watches.discard(uuid)
-
                 # Send delete signal
                 watch_delete_signal = signal('watch_deleted')
                 if watch_delete_signal:
                     watch_delete_signal.send(watch_uuid=uuid)
 
-        self.needs_write_urgent = True
-
     # Clone a watch by UUID
     def clone(self, uuid):
         url = self.data['watching'][uuid].get('url')
-        extras = deepcopy(self.data['watching'][uuid])
+        # No need to deepcopy here - add_watch() will deepcopy extras anyway (line 569)
+        # Just pass a dict copy (with lock for thread safety)
+        # NOTE: dict() is shallow copy but safe since add_watch() deepcopies it
+        with self.lock:
+            extras = dict(self.data['watching'][uuid])
         new_uuid = self.add_watch(url=url, extras=extras)
         watch = self.data['watching'][new_uuid]
         return new_uuid
@@ -544,7 +556,7 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
     # Remove a watchs data but keep the entry (URL etc)
     def clear_watch_history(self, uuid):
         self.__data['watching'][uuid].clear_watch()
-        self.needs_write_urgent = True
+        self.__data['watching'][uuid].commit()
 
     def add_watch(self, url, tag='', extras=None, tag_uuids=None, save_immediately=True):
 
@@ -639,7 +651,7 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
 
         # If the processor also has its own Watch implementation
         watch_class = get_custom_watch_obj_for_processor(apply_extras.get('processor'))
-        new_watch = watch_class(datastore_path=self.datastore_path, url=url)
+        new_watch = watch_class(datastore_path=self.datastore_path, __datastore=self.__data, url=url)
 
         new_uuid = new_watch.get('uuid')
 
@@ -657,16 +669,9 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
         self.__data['watching'][new_uuid] = new_watch
 
         if save_immediately:
-            # Save immediately using polymorphic method
-            try:
-                self.save_watch(new_uuid, force=True)
-                logger.debug(f"Saved new watch {new_uuid}")
-            except Exception as e:
-                logger.error(f"Failed to save new watch {new_uuid}: {e}")
-                # Mark dirty for retry
-                self.mark_watch_dirty(new_uuid)
-        else:
-            self.mark_watch_dirty(new_uuid)
+            # Save immediately using commit
+            new_watch.commit()
+            logger.debug(f"Saved new watch {new_uuid}")
 
         logger.debug(f"Added '{url}'")
 
@@ -858,16 +863,20 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
         # So we use the same model as a Watch
         with self.lock:
             from ..model import Tag
-            new_tag = Tag.model(datastore_path=self.datastore_path, default={
-                'title': title.strip(),
-                'date_created': int(time.time())
-            })
+            new_tag = Tag.model(
+                datastore_path=self.datastore_path,
+                __datastore=self.__data,
+                default={
+                    'title': title.strip(),
+                    'date_created': int(time.time())
+                }
+            )
 
             new_uuid = new_tag.get('uuid')
 
             self.__data['settings']['application']['tags'][new_uuid] = new_tag
 
-        self.mark_settings_dirty()
+        self.commit()
         return new_uuid
 
     def get_all_tags_for_watch(self, uuid):
@@ -984,7 +993,7 @@ class ChangeDetectionStore(DatastoreUpdatesMixin, FileSavingDataStore):
             notification_urls.append(notification_url)
             self.__data['settings']['application']['notification_urls'] = notification_urls
 
-        self.mark_settings_dirty()
+        self.commit()
         return notification_url
 
     # Schema update methods moved to store/updates.py (DatastoreUpdatesMixin)

changedetectionio/store/base.py

@@ -81,20 +81,3 @@ class DataStore(ABC):
         """
         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

@@ -1,22 +1,17 @@
 """
-File-based datastore with individual watch persistence and dirty tracking.
+File-based datastore with individual watch persistence and immediate commits.
 
 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
+- Immediate commit-based persistence (watch.commit(), datastore.commit())
 - 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
@@ -34,19 +29,6 @@ except ImportError:
 # 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
 # ============================================================================
@@ -61,6 +43,9 @@ def save_json_atomic(file_path, data_dict, label="file", max_size_mb=10):
     - Size validation
     - Proper error handling
 
+    Thread safety: Caller must hold datastore.lock to prevent concurrent modifications.
+    Multi-process safety: Not supported - run only one app instance per datastore.
+
     Args:
         file_path: Full path to target JSON file
         data_dict: Dictionary to serialize
@@ -242,11 +227,6 @@ def load_watch_from_file(watch_json, uuid, rehydrate_entity_func):
             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)
@@ -278,7 +258,7 @@ def load_watch_from_file(watch_json, uuid, rehydrate_entity_func):
         return None, None
 
 
-def load_all_watches(datastore_path, rehydrate_entity_func, compute_hash_func):
+def load_all_watches(datastore_path, rehydrate_entity_func):
     """
     Load all watches from individual watch.json files.
 
@@ -289,21 +269,17 @@ def load_all_watches(datastore_path, rehydrate_entity_func, compute_hash_func):
     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)
+        Dictionary of uuid -> Watch object
     """
     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
+        return watching
 
     # Find all watch.json files using glob (faster than manual directory traversal)
     glob_start = time.time()
@@ -322,9 +298,6 @@ def load_all_watches(datastore_path, rehydrate_entity_func, compute_hash_func):
         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:
@@ -344,7 +317,7 @@ def load_all_watches(datastore_path, rehydrate_entity_func, compute_hash_func):
     else:
         logger.info(f"Loaded {loaded} watches from disk in {elapsed:.2f}s ({loaded/elapsed:.0f} watches/sec)")
 
-    return watching, watch_hashes
+    return watching
 
 
 # ============================================================================
@@ -353,151 +326,20 @@ def load_all_watches(datastore_path, rehydrate_entity_func, compute_hash_func):
 
 class FileSavingDataStore(DataStore):
     """
-    Abstract datastore that provides file persistence with change tracking.
+    Abstract datastore that provides file persistence with immediate commits.
 
     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
+    - Immediate persistence via watch.commit() and datastore.commit()
+    - Atomic file writes for crash safety
 
     Subclasses must implement:
     - rehydrate_entity(): Convert dict to Watch object
     - Access to internal __data structure for watch management
     """
 
-    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):
         """
@@ -510,6 +352,7 @@ class FileSavingDataStore(DataStore):
         """
         raise NotImplementedError("Subclass must implement _save_settings")
 
+
     def _load_watches(self):
         """
         Load all watches from storage (polymorphic).
@@ -535,364 +378,4 @@ class FileSavingDataStore(DataStore):
         """
         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/updates.py

@@ -168,7 +168,7 @@ class DatastoreUpdatesMixin:
                     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()
+                    self.commit()
                     return  # No updates needed for fresh install
                 else:
                     # Has watches but no schema version - likely old datastore, run all updates
@@ -201,14 +201,14 @@ class DatastoreUpdatesMixin:
                 else:
                     # Bump the version, important
                     self.data['settings']['application']['schema_version'] = update_n
-                    self.mark_settings_dirty()
+                    self.commit()
 
-                    # CRITICAL: Mark all watches as dirty so changes are persisted
+                    # CRITICAL: Save all watches 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)")
+                    logger.info(f"Saving all {len(self.data['watching'])} watches after update_{update_n} (so that it saves them to disk)")
                     for uuid in self.data['watching'].keys():
-                        self.mark_watch_dirty(uuid)
+                        self.data['watching'][uuid].commit()
 
                     # Save changes immediately after each update (more resilient than batching)
                     logger.critical(f"Saving all changes after update_{update_n}")
@@ -662,7 +662,7 @@ class DatastoreUpdatesMixin:
         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()
+        self.commit()
         logger.info(f"Set schema_version to {latest_schema} (migration complete, all watches already saved)")
 
         logger.critical("=" * 80)

changedetectionio/tests/conftest.py

@@ -308,10 +308,6 @@ def prepare_test_function(live_server, datastore_path):
 
 
 
-    # Prevent background thread from writing during cleanup/reload
-    datastore.needs_write = False
-    datastore.needs_write_urgent = False
-
     # CRITICAL: Clean up any files from previous tests
     # This ensures a completely clean directory
     cleanup(datastore_path)
@@ -344,7 +340,6 @@ def prepare_test_function(live_server, datastore_path):
                 break
 
         datastore.data['watching'] = {}
-        datastore.needs_write = True
     except Exception as e:
         logger.warning(f"Error during datastore cleanup: {e}")
 

changedetectionio/tests/plugins/test_processor.py

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

changedetectionio/tests/test_api.py

@@ -465,7 +465,10 @@ def test_api_watch_PUT_update(client, live_server, measure_memory_usage, datasto
 
     assert res.status_code == 400, "Should get error 400 when we give a field that doesnt exist"
     # Message will come from `flask_expects_json`
-    assert b'Additional properties are not allowed' in res.data
+    # With patternProperties for processor_config_*, the error message format changed slightly
+    assert (b'Additional properties are not allowed' in res.data or
+            b'does not match any of the regexes' in res.data), \
+            "Should reject unknown fields with schema validation error"
 
 
     # Try a XSS URL

changedetectionio/tests/test_api_openapi.py

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

changedetectionio/tests/test_commit_persistence.py

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

changedetectionio/tests/test_history_consistency.py

@@ -182,3 +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/unit/test_watch_model.py

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

changedetectionio/tests/util.py

@@ -161,11 +161,6 @@ def extract_UUID_from_client(client):
 
 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)

changedetectionio/translations/cs/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2026-01-22 06:19+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: 2026-01-02 11:40+0100\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language: cs\n"
@@ -327,6 +327,14 @@ msgstr "Nastavit na"
 msgid "to disable"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr ""
@@ -345,6 +353,10 @@ msgstr ""
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr "Vyberte výchozí proxy pro všechny monitory"
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr ""
@@ -654,10 +666,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr ""
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr "Vyberte výchozí proxy pro všechny monitory"
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr "Verze Pythonu:"
@@ -983,7 +991,12 @@ msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
 msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
@@ -1232,6 +1245,10 @@ msgid ""
 "your filter will not work anymore."
 msgstr ""
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr ""
@@ -1580,6 +1597,18 @@ msgstr "zobrazeno <b>{start} - {end}</b> {record_name} z celkem <b>{total}</b>"
 msgid "records"
 msgstr "záznamy"
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr "Více informací"
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr "Přidejte nové monitory zjišťování změn webové stránky"
@@ -1592,18 +1621,6 @@ msgstr "Monitorovat tuto URL!"
 msgid "Edit first then Watch"
 msgstr "Upravit a monitorovat"
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr "Vytvořte odkaz ke sdílení"
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr "Tip: Můžete také přidat „sdílené“ monitory."
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr "Více informací"
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr "Pauza"
@@ -2117,6 +2134,10 @@ msgstr "Přiřaďte kteroukoli z následujících možností"
 msgid "Use page <title> in list"
 msgstr "V seznamu použijte stránku <title>"
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr "Když je metoda požadavku nastavena na GET, tělo musí být prázdné"
@@ -2439,15 +2460,20 @@ msgstr "Změny textu webové stránky/HTML, JSON a PDF"
 msgid "Detects all text changes where possible"
 msgstr "Detekuje všechny změny textu, kde je to možné"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr ""
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr ""
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr "Tělo pro všechna oznámení — Můžete použít"
@@ -3056,3 +3082,12 @@ msgstr "Hlavní nastavení"
 #~ msgid "Cleared snapshot history for all watches"
 #~ msgstr "Vymazat/resetovat historii"
 
+#~ msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+#~ msgstr ""
+
+#~ msgid "Create a shareable link"
+#~ msgstr "Vytvořte odkaz ke sdílení"
+
+#~ msgid "Tip: You can also add 'shared' watches."
+#~ msgstr "Tip: Můžete také přidat „sdílené“ monitory."
+

changedetectionio/translations/de/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2026-01-22 06:19+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: 2026-01-14 03:57+0100\n"
 "Last-Translator: \n"
 "Language: de\n"
@@ -333,6 +333,14 @@ msgstr "Setzen auf"
 msgid "to disable"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr ""
@@ -351,6 +359,10 @@ msgstr ""
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr "Wählen Sie einen Standard-Proxy für alle Überwachungen"
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr ""
@@ -664,10 +676,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr ""
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr "Wählen Sie einen Standard-Proxy für alle Überwachungen"
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr "Python-Version:"
@@ -999,8 +1007,13 @@ msgstr "In den Modus {} gewechselt."
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
-msgstr "Das Bearbeitungsformular für den Prozessor/das Plugin „{}“ kann nicht geladen werden. Fehlt das Plugin?"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
+msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 msgid "Updated watch - unpaused!"
@@ -1254,6 +1267,10 @@ msgstr ""
 "Sendet eine Benachrichtigung, wenn der Filter auf der Seite nicht mehr sichtbar ist. So wissen Sie, wann sich die "
 "Seite geändert hat und Ihr Filter nicht mehr funktioniert."
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr "Methode (default), bei der Ihre überwachte Website kein Javascript zum Rendern benötigt."
@@ -1618,6 +1635,18 @@ msgstr "zeige <b>{start} - {end}</b> {record_name} von insgesamt <b>{total}</b>"
 msgid "records"
 msgstr "Einträge"
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr "Weitere Informationen"
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr "Fügen Sie eine neue Überwachung zur Erkennung von Webseitenänderungen hinzu"
@@ -1630,18 +1659,6 @@ msgstr "Diese URL überwachen!"
 msgid "Edit first then Watch"
 msgstr "Bearbeiten > Überwachen"
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr "Erstellen Sie einen Link zum Teilen"
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr "Tipp: Sie können auch „gemeinsame“ Überwachungen hinzufügen."
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr "Weitere Informationen"
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr "Pause"
@@ -2162,6 +2179,10 @@ msgstr "Entspricht einer der folgenden Bedingungen"
 msgid "Use page <title> in list"
 msgstr "Verwenden Sie Seite <Titel> in der Liste"
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr "Der Textkörper muss leer sein, wenn die Anforderungsmethode auf GET gesetzt ist"
@@ -2486,15 +2507,20 @@ msgstr "Änderungen an Webseitentext/HTML, JSON und PDF"
 msgid "Detects all text changes where possible"
 msgstr "Erkennt nach Möglichkeit alle Textänderungen"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr "Fehler beim Abrufen der Metadaten für {}"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr "Das Protokoll wird nicht unterstützt oder das URL-Format ist ungültig."
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr "Inhalt für alle Benachrichtigungen — Sie können verwenden"
@@ -3171,3 +3197,12 @@ msgstr "Haupteinstellungen"
 #~ msgid "No watches available to recheck."
 #~ msgstr "Keine Überwachungen verfügbar, um erneut zu überprüfen."
 
+#~ msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+#~ msgstr "Das Bearbeitungsformular für den Prozessor/das Plugin „{}“ kann nicht geladen werden. Fehlt das Plugin?"
+
+#~ msgid "Create a shareable link"
+#~ msgstr "Erstellen Sie einen Link zum Teilen"
+
+#~ msgid "Tip: You can also add 'shared' watches."
+#~ msgstr "Tipp: Sie können auch „gemeinsame“ Überwachungen hinzufügen."
+

changedetectionio/translations/en_GB/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version:  changedetection.io\n"
 "Report-Msgid-Bugs-To: https://github.com/dgtlmoon/changedetection.io\n"
-"POT-Creation-Date: 2026-01-22 06:19+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: 2026-01-12 16:33+0100\n"
 "Last-Translator: British English Translation Team\n"
 "Language: en_GB\n"
@@ -325,6 +325,14 @@ msgstr ""
 msgid "to disable"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr ""
@@ -341,6 +349,10 @@ msgstr ""
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr ""
@@ -650,10 +662,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr ""
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr ""
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr ""
@@ -979,7 +987,12 @@ msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
 msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
@@ -1228,6 +1241,10 @@ msgid ""
 "your filter will not work anymore."
 msgstr ""
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr ""
@@ -1576,6 +1593,18 @@ msgstr ""
 msgid "records"
 msgstr ""
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr ""
@@ -1588,18 +1617,6 @@ msgstr ""
 msgid "Edit first then Watch"
 msgstr ""
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr ""
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr ""
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr ""
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr ""
@@ -2113,6 +2130,10 @@ msgstr ""
 msgid "Use page <title> in list"
 msgstr ""
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr ""
@@ -2435,15 +2456,20 @@ msgstr ""
 msgid "Detects all text changes where possible"
 msgstr ""
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr ""
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr ""
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr ""
@@ -3001,3 +3027,12 @@ msgstr ""
 #~ msgid "No watches available to recheck."
 #~ msgstr ""
 
+#~ msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+#~ msgstr ""
+
+#~ msgid "Create a shareable link"
+#~ msgstr ""
+
+#~ msgid "Tip: You can also add 'shared' watches."
+#~ msgstr ""
+

changedetectionio/translations/en_US/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: https://github.com/dgtlmoon/changedetection.io\n"
-"POT-Creation-Date: 2026-01-22 06:19+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: 2026-01-12 16:37+0100\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language: en_US\n"
@@ -325,6 +325,14 @@ msgstr ""
 msgid "to disable"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr ""
@@ -341,6 +349,10 @@ msgstr ""
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr ""
@@ -650,10 +662,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr ""
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr ""
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr ""
@@ -979,7 +987,12 @@ msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
 msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
@@ -1228,6 +1241,10 @@ msgid ""
 "your filter will not work anymore."
 msgstr ""
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr ""
@@ -1576,6 +1593,18 @@ msgstr ""
 msgid "records"
 msgstr ""
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr ""
@@ -1588,18 +1617,6 @@ msgstr ""
 msgid "Edit first then Watch"
 msgstr ""
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr ""
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr ""
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr ""
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr ""
@@ -2113,6 +2130,10 @@ msgstr ""
 msgid "Use page <title> in list"
 msgstr ""
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr ""
@@ -2435,15 +2456,20 @@ msgstr ""
 msgid "Detects all text changes where possible"
 msgstr ""
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr ""
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr ""
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr ""
@@ -3001,3 +3027,12 @@ msgstr ""
 #~ msgid "No watches available to recheck."
 #~ msgstr ""
 
+#~ msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+#~ msgstr ""
+
+#~ msgid "Create a shareable link"
+#~ msgstr ""
+
+#~ msgid "Tip: You can also add 'shared' watches."
+#~ msgstr ""
+

changedetectionio/translations/fr/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2026-01-22 06:19+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: 2026-01-02 11:40+0100\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language: fr\n"
@@ -327,6 +327,14 @@ msgstr "Définir à"
 msgid "to disable"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr ""
@@ -345,6 +353,10 @@ msgstr ""
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr "Choisir un proxy par défaut pour tous les moniteurs"
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr ""
@@ -654,10 +666,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr ""
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr "Choisir un proxy par défaut pour tous les moniteurs"
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr "Version Python :"
@@ -983,7 +991,12 @@ msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
 msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
@@ -1234,6 +1247,10 @@ msgid ""
 "your filter will not work anymore."
 msgstr ""
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr ""
@@ -1582,6 +1599,18 @@ msgstr "affichage de <b>{start} - {end}</b> {record_name} sur un total de <b>{to
 msgid "records"
 msgstr "enregistrements"
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr "Plus d'informations"
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr "Ajouter une nouvelle surveillance de détection de changement de page Web"
@@ -1594,18 +1623,6 @@ msgstr "Surveillez cette URL !"
 msgid "Edit first then Watch"
 msgstr "Modifier > Surveiller"
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr "Créer un lien partageable"
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr "Astuce : Vous pouvez également ajouter des montres « partagées »."
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr "Plus d'informations"
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr "Pause"
@@ -2123,6 +2140,10 @@ msgstr "Faites correspondre l'un des éléments suivants"
 msgid "Use page <title> in list"
 msgstr "Utiliser la page <titre> dans la liste"
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr "Le corps doit être vide lorsque la méthode de requête est définie sur GET"
@@ -2445,15 +2466,20 @@ msgstr "Modifications du texte de la page Web/HTML, JSON et PDF"
 msgid "Detects all text changes where possible"
 msgstr "Détecte toutes les modifications de texte lorsque cela est possible"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr ""
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr ""
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr "Corps pour toutes les notifications — Vous pouvez utiliser"
@@ -3064,3 +3090,12 @@ msgstr "Paramètres principaux"
 #~ msgid "Cleared snapshot history for all watches"
 #~ msgstr "Effacer/réinitialiser l'historique"
 
+#~ msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+#~ msgstr ""
+
+#~ msgid "Create a shareable link"
+#~ msgstr "Créer un lien partageable"
+
+#~ msgid "Tip: You can also add 'shared' watches."
+#~ msgstr "Astuce : Vous pouvez également ajouter des montres « partagées »."
+

changedetectionio/translations/it/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2026-01-22 06:19+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: 2026-01-02 15:32+0100\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language: it\n"
@@ -327,6 +327,14 @@ msgstr ""
 msgid "to disable"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr ""
@@ -343,6 +351,10 @@ msgstr "Consenti accesso alla pagina cronologia quando la password è attiva (ut
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr ""
@@ -652,10 +664,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr ""
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr ""
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr ""
@@ -981,7 +989,12 @@ msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
 msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
@@ -1230,6 +1243,10 @@ msgid ""
 "your filter will not work anymore."
 msgstr ""
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr ""
@@ -1578,6 +1595,18 @@ msgstr "visualizzando <b>{start} - {end}</b> {record_name} su un totale di <b>{t
 msgid "records"
 msgstr "record"
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr "Maggiori informazioni"
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr "Aggiungi un nuovo monitoraggio modifiche pagina web"
@@ -1590,18 +1619,6 @@ msgstr "Monitora questo URL!"
 msgid "Edit first then Watch"
 msgstr "Modifica > Monitora"
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr ""
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr ""
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr "Maggiori informazioni"
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr "Pausa"
@@ -2115,6 +2132,10 @@ msgstr "Corrisponde a uno qualsiasi dei seguenti"
 msgid "Use page <title> in list"
 msgstr "Usa <title> pagina nell'elenco"
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr "Il corpo deve essere vuoto quando il metodo è impostato su GET"
@@ -2437,15 +2458,20 @@ msgstr "Modifiche testo/HTML, JSON e PDF"
 msgid "Detects all text changes where possible"
 msgstr "Rileva tutte le modifiche di testo possibili"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr "Errore nel recupero metadati per {}"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr "Protocollo non consentito o formato URL non valido"
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr "Corpo per tutte le notifiche — Puoi usare"
@@ -3036,3 +3062,12 @@ msgstr "Impostazioni principali"
 #~ msgid "Queue"
 #~ msgstr "In coda"
 
+#~ msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+#~ msgstr ""
+
+#~ msgid "Create a shareable link"
+#~ msgstr ""
+
+#~ msgid "Tip: You can also add 'shared' watches."
+#~ msgstr ""
+

changedetectionio/translations/ko/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2026-01-22 06:19+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: 2026-01-02 11:40+0100\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language: ko\n"
@@ -325,6 +325,14 @@ msgstr "설정:"
 msgid "to disable"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr ""
@@ -341,6 +349,10 @@ msgstr "비밀번호 활성화 시 변경 기록 페이지 액세스 허용 (dif
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr "모든 모니터의 기본 프록시 선택"
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr ""
@@ -650,10 +662,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr ""
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr "모든 모니터의 기본 프록시 선택"
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr "파이썬 버전:"
@@ -979,7 +987,12 @@ msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
 msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
@@ -1228,6 +1241,10 @@ msgid ""
 "your filter will not work anymore."
 msgstr ""
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr ""
@@ -1576,6 +1593,18 @@ msgstr "총 <b>{total}</b>개 중 <b>{start} - {end}</b>개 {record_name} 표시
 msgid "records"
 msgstr "기록"
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr "추가 정보"
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr "새로운 웹 페이지 변경 감지 감시 추가"
@@ -1588,18 +1617,6 @@ msgstr "이 URL 모니터!"
 msgid "Edit first then Watch"
 msgstr "편집 후 모니터"
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr "공유 가능한 링크 만들기"
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr "팁: '공유' 시계를 추가할 수도 있습니다."
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr "추가 정보"
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr "정지시키다"
@@ -2113,6 +2130,10 @@ msgstr "다음 중 하나와 일치"
 msgid "Use page <title> in list"
 msgstr "목록의 <제목> 페이지 사용"
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr "요청 방법이 GET으로 설정된 경우 본문이 비어 있어야 합니다."
@@ -2435,15 +2456,20 @@ msgstr "웹페이지 텍스트/HTML, JSON 및 PDF 변경"
 msgid "Detects all text changes where possible"
 msgstr "가능한 경우 모든 텍스트 변경 사항을 감지합니다."
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr ""
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr ""
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr "모든 알림 본문 — 사용 가능:"
@@ -3157,3 +3183,12 @@ msgstr "기본 설정"
 #~ msgid "Cleared snapshot history for all watches"
 #~ msgstr "기록 지우기/재설정"
 
+#~ msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+#~ msgstr ""
+
+#~ msgid "Create a shareable link"
+#~ msgstr "공유 가능한 링크 만들기"
+
+#~ msgid "Tip: You can also add 'shared' watches."
+#~ msgstr "팁: '공유' 시계를 추가할 수도 있습니다."
+

changedetectionio/translations/messages.pot

@@ -6,9 +6,9 @@
 #, fuzzy
 msgid ""
 msgstr ""
-"Project-Id-Version: changedetection.io 0.52.8\n"
+"Project-Id-Version: changedetection.io 0.52.9\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2026-01-22 06:29+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language-Team: LANGUAGE <LL@li.org>\n"
@@ -324,6 +324,14 @@ msgstr ""
 msgid "to disable"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr ""
@@ -340,6 +348,10 @@ msgstr ""
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr ""
@@ -649,10 +661,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr ""
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr ""
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr ""
@@ -978,7 +986,12 @@ msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
 msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
@@ -1227,6 +1240,10 @@ msgid ""
 "your filter will not work anymore."
 msgstr ""
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr ""
@@ -1575,6 +1592,18 @@ msgstr ""
 msgid "records"
 msgstr ""
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr ""
@@ -1587,18 +1616,6 @@ msgstr ""
 msgid "Edit first then Watch"
 msgstr ""
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr ""
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr ""
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr ""
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr ""
@@ -2112,6 +2129,10 @@ msgstr ""
 msgid "Use page <title> in list"
 msgstr ""
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr ""
@@ -2434,15 +2455,20 @@ msgstr ""
 msgid "Detects all text changes where possible"
 msgstr ""
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr ""
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr ""
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr ""

changedetectionio/translations/zh/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2026-01-22 06:19+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: 2026-01-18 21:31+0800\n"
 "Last-Translator: 吾爱分享 <admin@wuaishare.cn>\n"
 "Language: zh\n"
@@ -325,6 +325,14 @@ msgstr "设置为"
 msgid "to disable"
 msgstr "以禁用"
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr "为你的 changedetection.io 应用启用密码保护。"
@@ -341,6 +349,10 @@ msgstr "启用密码时允许访问监视器更改历史页面(便于共享差
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr "当请求无内容返回，或 HTML 不包含任何文本时，是否视为变更？"
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr "为所有监视器选择默认代理"
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr "用于通知链接中的"
@@ -650,10 +662,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr "带认证的 SOCKS5 代理仅支持“明文请求”抓取器，其他抓取器请改为白名单 IP"
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr "为所有监视器选择默认代理"
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr "Python 版本："
@@ -979,8 +987,13 @@ msgstr "已切换到模式 - {}。"
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
-msgstr "无法加载处理器/插件 '{}' 的编辑表单，插件是否缺失？"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
+msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 msgid "Updated watch - unpaused!"
@@ -1228,6 +1241,10 @@ msgid ""
 "your filter will not work anymore."
 msgstr "当页面上找不到该过滤器时发送通知，便于知晓页面已变化且过滤器不再适用。"
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr "方式（默认），适用于无需 JavaScript 渲染的网站。"
@@ -1576,6 +1593,18 @@ msgstr "显示第 <b>{start} - {end}</b> 条{record_name}，共 <b>{total}</b>
 msgid "records"
 msgstr "记录"
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr "更多信息"
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr "新增网页变更监控"
@@ -1588,18 +1617,6 @@ msgstr "监控此 URL！"
 msgid "Edit first then Watch"
 msgstr "先编辑，再监控"
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr "创建可分享链接"
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr "提示：你也可以添加“共享”的监控项。"
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr "更多信息"
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr "暂停"
@@ -2113,6 +2130,10 @@ msgstr "匹配以下任意"
 msgid "Use page <title> in list"
 msgstr "列表中使用页面 <title>"
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr "当请求方法为 GET 时，请求正文必须为空"
@@ -2435,15 +2456,20 @@ msgstr "网页文本/HTML、JSON 和 PDF 变更"
 msgid "Detects all text changes where possible"
 msgstr "尽可能检测所有文本变更"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr "获取 {} 的元数据失败"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr "监控协议不允许或 URL 格式无效"
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr "所有通知的正文 — 您可以使用"
@@ -2986,3 +3012,12 @@ msgstr "否"
 msgid "Main settings"
 msgstr "主设置"
 
+#~ msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+#~ msgstr "无法加载处理器/插件 '{}' 的编辑表单，插件是否缺失？"
+
+#~ msgid "Create a shareable link"
+#~ msgstr "创建可分享链接"
+
+#~ msgid "Tip: You can also add 'shared' watches."
+#~ msgstr "提示：你也可以添加“共享”的监控项。"
+

changedetectionio/translations/zh_Hant_TW/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2026-01-22 06:19+0100\n"
+"POT-Creation-Date: 2026-02-05 17:47+0100\n"
 "PO-Revision-Date: 2026-01-15 12:00+0800\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language: zh_Hant_TW\n"
@@ -325,6 +325,14 @@ msgstr "設置為"
 msgid "to disable"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html changedetectionio/blueprint/ui/templates/edit.html
+msgid "Limit collection of history snapshots for each watch to this number of history items."
+msgstr ""
+
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Set to empty to disable / no limit"
+msgstr ""
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Password protection for your changedetection.io application."
 msgstr ""
@@ -341,6 +349,10 @@ msgstr "啟用密碼時允許匿名存取監測歷史頁面"
 msgid "When a request returns no content, or the HTML does not contain any text, is this considered a change?"
 msgstr ""
 
+#: changedetectionio/blueprint/settings/templates/settings.html
+msgid "Choose a default proxy for all watches"
+msgstr "為所有監測任務選擇預設代理伺服器"
+
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Base URL used for the"
 msgstr ""
@@ -650,10 +662,6 @@ msgid ""
 "whitelist the IP access instead"
 msgstr ""
 
-#: changedetectionio/blueprint/settings/templates/settings.html
-msgid "Choose a default proxy for all watches"
-msgstr "為所有監測任務選擇預設代理伺服器"
-
 #: changedetectionio/blueprint/settings/templates/settings.html
 msgid "Python version:"
 msgstr "Python 版本："
@@ -979,8 +987,13 @@ msgstr "已切換至模式 - {}。"
 
 #: changedetectionio/blueprint/ui/edit.py
 #, python-brace-format
-msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
-msgstr "無法載入處理器 / 外掛 '{}' 的編輯表單，外掛是否遺失？"
+msgid "Could not load '{}' processor, processor plugin might be missing. Please select a different processor."
+msgstr ""
+
+#: changedetectionio/blueprint/ui/edit.py
+#, python-brace-format
+msgid "Could not load '{}' processor, processor plugin might be missing."
+msgstr ""
 
 #: changedetectionio/blueprint/ui/edit.py
 msgid "Updated watch - unpaused!"
@@ -1228,6 +1241,10 @@ msgid ""
 "your filter will not work anymore."
 msgstr "當頁面上找不到過濾器時發送通知，這有助於了解頁面何時變更導致您的過濾器失效。"
 
+#: changedetectionio/blueprint/ui/templates/edit.html
+msgid "Set to empty to use system settings default"
+msgstr ""
+
 #: changedetectionio/blueprint/ui/templates/edit.html
 msgid "method (default) where your watched site doesn't need Javascript to render."
 msgstr "方法（預設），適用於您監測的網站不需要 Javascript 渲染的情況。"
@@ -1576,6 +1593,18 @@ msgstr "顯示第 <b>{start} - {end}</b> 條{record_name}，共 <b>{total}</b>
 msgid "records"
 msgstr "記錄"
 
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "Changedetection.io can monitor more than just web-pages! See our plugins!"
+msgstr ""
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "More info"
+msgstr "更多資訊"
+
+#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
+msgid "You can also add 'shared' watches."
+msgstr ""
+
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Add a new web page change detection watch"
 msgstr "新增網頁變更檢測任務"
@@ -1588,18 +1617,6 @@ msgstr "監測此 URL！"
 msgid "Edit first then Watch"
 msgstr "先編輯後監測"
 
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Create a shareable link"
-msgstr "建立可分享連結"
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "Tip: You can also add 'shared' watches."
-msgstr "提示：您也可以新增「共享」監測任務。"
-
-#: changedetectionio/blueprint/watchlist/templates/watch-overview.html
-msgid "More info"
-msgstr "更多資訊"
-
 #: changedetectionio/blueprint/watchlist/templates/watch-overview.html
 msgid "Pause"
 msgstr "暫停"
@@ -2113,6 +2130,10 @@ msgstr "符合以下任一條件"
 msgid "Use page <title> in list"
 msgstr "在列表中使用頁面 <title>"
 
+#: changedetectionio/forms.py
+msgid "Number of history items per watch to keep"
+msgstr ""
+
 #: changedetectionio/forms.py
 msgid "Body must be empty when Request Method is set to GET"
 msgstr "當請求方法設為 GET 時，內容必須為空"
@@ -2435,15 +2456,20 @@ msgstr "網頁文字 / HTML、JSON 和 PDF 變更"
 msgid "Detects all text changes where possible"
 msgstr "盡可能檢測所有文字變更"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 #, python-brace-format
 msgid "Error fetching metadata for {}"
 msgstr "讀取 {} 的中繼資料時發生錯誤"
 
-#: changedetectionio/store.py
+#: changedetectionio/store/__init__.py
 msgid "Watch protocol is not permitted or invalid URL format"
 msgstr "監測協定不被允許或 URL 格式無效"
 
+#: changedetectionio/store/__init__.py
+#, python-brace-format
+msgid "Watch limit reached ({}/{} watches). Cannot add more watches."
+msgstr ""
+
 #: changedetectionio/templates/_common_fields.html
 msgid "Body for all notifications — You can use"
 msgstr "所有通知的內文 — 您可以使用"
@@ -3115,3 +3141,12 @@ msgstr "主設定"
 #~ msgid "No watches available to recheck."
 #~ msgstr "沒有可複查的監測任務。"
 
+#~ msgid "Cannot load the edit form for processor/plugin '{}', plugin missing?"
+#~ msgstr "無法載入處理器 / 外掛 '{}' 的編輯表單，外掛是否遺失？"
+
+#~ msgid "Create a shareable link"
+#~ msgstr "建立可分享連結"
+
+#~ msgid "Tip: You can also add 'shared' watches."
+#~ msgstr "提示：您也可以新增「共享」監測任務。"
+

changedetectionio/worker.py

@@ -142,11 +142,14 @@ async def async_update_worker(worker_id, q, notification_q, app, datastore, exec
                     processor = watch.get('processor', 'text_json_diff')
 
                     # Init a new 'difference_detection_processor'
-                    try:
-                        processor_module = importlib.import_module(f"changedetectionio.processors.{processor}.processor")
-                    except ModuleNotFoundError as e:
-                        print(f"Processor module '{processor}' not found.")
-                        raise e
+                    # Use get_processor_module() to support both built-in and plugin processors
+                    from changedetectionio.processors import get_processor_module
+                    processor_module = get_processor_module(processor)
+
+                    if not processor_module:
+                        error_msg = f"Processor module '{processor}' not found."
+                        logger.error(error_msg)
+                        raise ModuleNotFoundError(error_msg)
 
                     update_handler = processor_module.perform_site_check(datastore=datastore,
                                                                          watch_uuid=uuid)
@@ -365,8 +368,9 @@ async def async_update_worker(worker_id, q, notification_q, app, datastore, exec
                     logger.error(f"Exception (BrowserStepsInUnsupportedFetcher) reached processing watch UUID: {uuid}")
 
                 except Exception as e:
+                    import traceback
                     logger.error(f"Worker {worker_id} exception processing watch UUID: {uuid}")
-                    logger.error(str(e))
+                    logger.exception(f"Worker {worker_id} full exception details:")
                     datastore.update_watch(uuid=uuid, update_obj={'last_error': "Exception: " + str(e)})
                     process_changedetection_results = False
 
@@ -385,8 +389,8 @@ async def async_update_worker(worker_id, q, notification_q, app, datastore, exec
                 if not datastore.data['watching'].get(uuid):
                     continue
 
-                logger.debug(f"Processing watch UUID: {uuid} - xpath_data length returned {len(update_handler.xpath_data) if update_handler.xpath_data else 'empty.'}")
-                if process_changedetection_results:
+                logger.debug(f"Processing watch UUID: {uuid} - xpath_data length returned {len(update_handler.xpath_data) if update_handler and update_handler.xpath_data else 'empty.'}")
+                if update_handler and process_changedetection_results:
                     try:
                         datastore.update_watch(uuid=uuid, update_obj=update_obj)
 
@@ -430,64 +434,62 @@ async def async_update_worker(worker_id, q, notification_q, app, datastore, exec
                                     await send_content_changed_notification(uuid, notification_q, datastore)
 
                     except Exception as e:
+
                         logger.critical(f"Worker {worker_id} exception in process_changedetection_results")
-                        logger.critical(str(e))
+                        logger.exception(f"Worker {worker_id} full exception details:")
                         datastore.update_watch(uuid=uuid, update_obj={'last_error': str(e)})
 
                 # Always record attempt count
                 count = watch.get('check_count', 0) + 1
+                if update_handler: # Could be none or empty if the processor was not found
+                    # Always record page title (used in notifications, and can change even when the content is the same)
+                    if update_obj.get('content-type') and 'html' in update_obj.get('content-type'):
+                        try:
+                            page_title = html_tools.extract_title(data=update_handler.fetcher.content)
+                            if page_title:
+                                page_title = page_title.strip()[:2000]
+                                logger.debug(f"UUID: {uuid} Page <title> is '{page_title}'")
+                                datastore.update_watch(uuid=uuid, update_obj={'page_title': page_title})
+                        except Exception as e:
+                            logger.exception(f"Worker {worker_id} full exception details:")
+                            logger.warning(f"UUID: {uuid} Exception when extracting <title> - {str(e)}")
 
-                # Always record page title (used in notifications, and can change even when the content is the same)
-                if update_obj.get('content-type') and 'html' in update_obj.get('content-type'):
+                    # Record server header
                     try:
-                        page_title = html_tools.extract_title(data=update_handler.fetcher.content)
-                        if page_title:
-                            page_title = page_title.strip()[:2000]
-                            logger.debug(f"UUID: {uuid} Page <title> is '{page_title}'")
-                            datastore.update_watch(uuid=uuid, update_obj={'page_title': page_title})
+                        server_header = update_handler.fetcher.headers.get('server', '').strip().lower()[:255]
+                        datastore.update_watch(uuid=uuid, update_obj={'remote_server_reply': server_header})
                     except Exception as e:
-                        logger.warning(f"UUID: {uuid} Exception when extracting <title> - {str(e)}")
+                        pass
 
-                # Record server header
-                try:
-                    server_header = update_handler.fetcher.headers.get('server', '').strip().lower()[:255]
-                    datastore.update_watch(uuid=uuid, update_obj={'remote_server_reply': server_header})
-                except Exception as e:
-                    pass
+                    # Store favicon if necessary
+                    if update_handler.fetcher.favicon_blob and update_handler.fetcher.favicon_blob.get('base64'):
+                        watch.bump_favicon(url=update_handler.fetcher.favicon_blob.get('url'),
+                                           favicon_base_64=update_handler.fetcher.favicon_blob.get('base64')
+                                           )
 
-                # Store favicon if necessary
-                if update_handler.fetcher.favicon_blob and update_handler.fetcher.favicon_blob.get('base64'):
-                    watch.bump_favicon(url=update_handler.fetcher.favicon_blob.get('url'),
-                                       favicon_base_64=update_handler.fetcher.favicon_blob.get('base64')
-                                       )
+                    datastore.update_watch(uuid=uuid, update_obj={'fetch_time': round(time.time() - fetch_start_time, 3),
+                                                                   'check_count': count})
 
-                datastore.update_watch(uuid=uuid, update_obj={'fetch_time': round(time.time() - fetch_start_time, 3),
-                                                               'check_count': count})
+                    # NOW clear fetcher content - after all processing is complete
+                    # This is the last point where we need the fetcher data
+                    if update_handler and hasattr(update_handler, 'fetcher') and update_handler.fetcher:
+                        update_handler.fetcher.clear_content()
+                        logger.debug(f"Cleared fetcher content for UUID {uuid}")
 
-                # NOW clear fetcher content - after all processing is complete
-                # This is the last point where we need the fetcher data
-                if update_handler and hasattr(update_handler, 'fetcher') and update_handler.fetcher:
-                    update_handler.fetcher.clear_content()
-                    logger.debug(f"Cleared fetcher content for UUID {uuid}")
+                    # Explicitly delete update_handler to free all references
+                    if update_handler:
+                        del update_handler
+                        update_handler = None
 
-                # Explicitly delete update_handler to free all references
-                if update_handler:
-                    del update_handler
-                    update_handler = None
-
-                # Force aggressive memory cleanup after clearing
+                # Force garbage collection
                 import gc
                 gc.collect()
-                try:
-                    import ctypes
-                    ctypes.CDLL('libc.so.6').malloc_trim(0)
-                except Exception:
-                    pass
 
         except Exception as e:
+
             logger.error(f"Worker {worker_id} unexpected error processing {uuid}: {e}")
-            logger.error(f"Worker {worker_id} traceback:", exc_info=True)
-            
+            logger.exception(f"Worker {worker_id} full exception details:")
+
             # Also update the watch with error information
             if datastore and uuid in datastore.data['watching']:
                 datastore.update_watch(uuid=uuid, update_obj={'last_error': f"Worker error: {str(e)}"})
@@ -495,49 +497,43 @@ async def async_update_worker(worker_id, q, notification_q, app, datastore, exec
         finally:
             # Always cleanup - this runs whether there was an exception or not
             if uuid:
+                # Call quit() as backup (Puppeteer/Playwright have internal cleanup, but this acts as safety net)
                 try:
                     if update_handler and hasattr(update_handler, 'fetcher') and update_handler.fetcher:
                         await update_handler.fetcher.quit(watch=watch)
                 except Exception as e:
                     logger.error(f"Exception while cleaning/quit after calling browser: {e}")
+                    logger.exception(f"Worker {worker_id} full exception details:")
+
                 try:
                     # Release UUID from processing (thread-safe)
                     worker_pool.release_uuid_from_processing(uuid, worker_id=worker_id)
-                    
+
                     # Send completion signal
                     if watch:
-                        #logger.info(f"Worker {worker_id} sending completion signal for UUID {watch['uuid']}")
                         watch_check_update.send(watch_uuid=watch['uuid'])
 
-                    # Explicitly clean up update_handler and all its references
+                    # Clean up all memory references BEFORE garbage collection
                     if update_handler:
-                        # Clear fetcher content using the proper method
                         if hasattr(update_handler, 'fetcher') and update_handler.fetcher:
                             update_handler.fetcher.clear_content()
-
-                        # Clear processor references
                         if hasattr(update_handler, 'content_processor'):
                             update_handler.content_processor = None
-
+                        del update_handler
                         update_handler = None
 
-                    # Clear local contents variable if it still exists
+                    # Clear large content variables
                     if 'contents' in locals():
                         del contents
 
-                    # Note: We don't set watch = None here because:
-                    # 1. watch is just a local reference to datastore.data['watching'][uuid]
-                    # 2. Setting it to None doesn't affect the datastore
-                    # 3. GC can't collect the object anyway (still referenced by datastore)
-                    # 4. It would just cause confusion
-
-                    # Force garbage collection after cleanup
+                    # Force garbage collection after all references are cleared
                     import gc
                     gc.collect()
 
                     logger.debug(f"Worker {worker_id} completed watch {uuid} in {time.time()-fetch_start_time:.2f}s")
                 except Exception as cleanup_error:
                     logger.error(f"Worker {worker_id} error during cleanup: {cleanup_error}")
+                    logger.exception(f"Worker {worker_id} full exception details:")
 
             del(uuid)
 

docker-compose.yml

@@ -16,6 +16,13 @@ services:
   #        Log output levels: TRACE, DEBUG(default), INFO, SUCCESS, WARNING, ERROR, CRITICAL
   #      - LOGGER_LEVEL=TRACE
   #
+  #        Plugins! See https://changedetection.io/plugins for more plugins.
+  #        Install additional Python packages (processor plugins, etc.)
+  #        Example: Install the OSINT reconnaissance processor plugin
+  #      - EXTRA_PACKAGES=changedetection.io-osint-processor
+  #        Multiple packages can be installed by separating with spaces:
+  #      - EXTRA_PACKAGES=changedetection.io-osint-processor another-plugin
+  #
   #
   #       Uncomment below and the "sockpuppetbrowser" to use a real Chrome browser (It uses the "playwright" protocol)
   #      - PLAYWRIGHT_DRIVER_URL=ws://browser-sockpuppet-chrome:3000

docker-entrypoint.sh

@@ -0,0 +1,28 @@
+#!/bin/bash
+set -e
+
+# Install additional packages from EXTRA_PACKAGES env var
+# Uses a marker file to avoid reinstalling on every container restart
+INSTALLED_MARKER="/datastore/.extra_packages_installed"
+CURRENT_PACKAGES="$EXTRA_PACKAGES"
+
+if [ -n "$EXTRA_PACKAGES" ]; then
+    # Check if we need to install/update packages
+    if [ ! -f "$INSTALLED_MARKER" ] || [ "$(cat $INSTALLED_MARKER 2>/dev/null)" != "$CURRENT_PACKAGES" ]; then
+        echo "Installing extra packages: $EXTRA_PACKAGES"
+        pip3 install --no-cache-dir $EXTRA_PACKAGES
+
+        if [ $? -eq 0 ]; then
+            echo "$CURRENT_PACKAGES" > "$INSTALLED_MARKER"
+            echo "Extra packages installed successfully"
+        else
+            echo "ERROR: Failed to install extra packages"
+            exit 1
+        fi
+    else
+        echo "Extra packages already installed: $EXTRA_PACKAGES"
+    fi
+fi
+
+# Execute the main command
+exec "$@"

requirements.txt

@@ -51,9 +51,9 @@ linkify-it-py
 
 # - Needed for apprise/spush, and maybe others? hopefully doesnt trigger a rust compile.
 # - Requires extra wheel for rPi, adds build time for arm/v8 which is not in piwheels
-# Pinned to 43.0.1 for ARM compatibility (45.x may not have pre-built ARM wheels)
+# Pinned to 44.x for ARM compatibility and sslyze compatibility (sslyze requires <45) and (45.x may not have pre-built ARM wheels)
 # Also pinned because dependabot wants specific versions
-cryptography==46.0.3
+cryptography==44.0.0
 
 # apprise mqtt https://github.com/dgtlmoon/changedetection.io/issues/315
 # use any version other than 2.0.x due to https://github.com/eclipse/paho.mqtt.python/issues/814
@@ -70,7 +70,7 @@ lxml >=4.8.0,!=5.2.0,!=5.2.1,<7
 
 # XPath 2.0-3.1 support - 4.2.0 had issues, 4.1.5 stable
 # Consider updating to latest stable version periodically
-elementpath==5.1.0
+elementpath==5.1.1
 
 # For fast image comparison in screenshot change detection
 # opencv-python-headless is OPTIONAL (excluded from requirements.txt)
@@ -91,7 +91,7 @@ jq~=1.3; python_version >= "3.8" and sys_platform == "linux"
 
 # playwright is installed at Dockerfile build time because it's not available on all platforms
 
-pyppeteer-ng==2.0.0rc12
+pyppeteer-ng==2.0.0rc13
 pyppeteerstealth>=0.0.4
 
 # Include pytest, so if theres a support issue we can ask them to run these tests on their setup
@@ -148,7 +148,7 @@ tzdata
 pluggy ~= 1.6
 
 # Needed for testing, cross-platform for process and system monitoring
-psutil==7.2.1
+psutil==7.2.2
 
 ruff >= 0.11.2
 pre_commit >= 4.2.0