refactor: Clean up state management #729

castielwaverly · 2025-09-19T21:20:51Z

Description

This started with taking a stab at

Huntarr Stage Management doesn't work, or I don't understand it #717

I kept discovering code that simultaneously seemed to be causing the problem but also doing nothing. Turns out there's multiple unused legacy state management systems chugging along in the background of the application and one of them is interfering with the current per-instance system.

Instead of trying to prevent the old systems from interfering with the new system, I just got rid of the old systems.

Testing

Built an image using this branch.

git clone https://github.com/castielwaverly/Huntarr.io.git
cd Huntarr.io
git checkout fix/717-instance-state-management
docker build -t huntarr-dev .

Made a clone of my existing huntarr setup.

cd $MEDIA_SERVER_PATH
cp -r huntar huntarr-dev

vim docker-compose.yml
# make a copy of the huntarr block and re-map a few things
# e.g. volume pointed to `huntarr-dev/config` image and container name set to `huntarr-dev`

Ran the modified image instead of the production one.

docker compose stop huntarr
docker compose up huntarr-dev -d
docker logs --follow huntarr-dev

At this point, it's part of my bigger setup and I'm keeping track of logs about sonarr and radarr cycles (time to reset, tracked items, etc.). Also paying attention to the relevant tables:

sqlite3 huntarr-dev/config/huntarr.db

sqlite> select * from stateful_instance_locks;
...

sqlite> select * from stateful_processed_ids;
...

Summary of results

radarr tracked items not getting wiped after a week ✅
state checks are logged accurately ✅
cycle summaries are logged accurately ✅
fresh install starts up successfully ✅
fresh install stateful lock initializations:
- sonarr/Default initialized (based on default settings) ✅
- radarr/Default not initialized (based on default settings) ✅
- lidarr/Default initialized (based on default settings) ✅
- readarr/Default initialized (based on default settings) ✅
- whisparr/Default initialized (based on default settings) ✅
- eros/Default initialized (based on default settings) ✅
newly saved instances' stateful locks are initialized ✅
newly saved instance runs cycles correctly ✅

Opportunistic fixes & improvements

Fix state management summaries

State management summaries were broken; this whole block does not run because configured_instances doesn't exist. As a result, we always use defaults for a few details in the summary -- the reset interval being the most obvious.

Logs below with the error and the default interval, despite a 72h interval configured for sonar:

huntarr.sonarr - WARNING - Could not load instance settings for sonarr: name 'configured_instances' is not defined
...
huntarr.sonarr - INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
huntarr.sonarr - INFO -   sonarr: 42 items tracked, next reset: 2025-09-21 06:48:15 (168h interval)

Now:

huntarr.sonarr - INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
huntarr.sonarr - INFO -   sonarr: 42 items tracked, next reset: 2025-09-21 06:48:15 (72h interval)

Reduce reinitializations and reset checks

Currently state management is mostly handled within is_processed and add_processed_id. These are called a lot during each cycle, so you end up doing a settings load and running several queries many times unnecessarily during each cycle. The lifecycle of instances is defined: some exist on startup and the rest are added through settings. If you initialize state management at both these points, you can avoid redoing it many times every cycle and simplify the core loop of the application.

Cleanup of timezone logic

Looks like there was a partial centralization of get_user_timestamp logic to a shared place, but the original duplicated implementations were left as wrappers of the centralized logic. The wrappers were removed because they didn't differ in any way from the function they were wrapping.

Separation of concerns

In general, there are layers of the app doing the job of another's. The scope of the changes here are to limited to instance level state management, so only a subset of relevant parts were moved around. More specifically, some business logic was moved out of routes, background cycle coordination, and database layers and into business logic layers.

Dead code removal

There were a few orphaned functions and commented out code from previous refactors. Using the repo history to reference old code is much more sustainable, so that code was removed.

Alignment with PEP8

Python has long-standing, well-defined standards around whitespace, spacing, and imports; these should be followed to reduce friction during development. These were addressed (mostly automatically by an IDE) in the files that were touched.

References:

Lazy formatting in logging

It's generally a good practice to use lazy formatting when logging. While less common of a practice than following PEP8, it is still a common standard that ought to be aligned to unless there's good reason not to. Logging statements were converted (using automation) in the files that were touched.

References:

castielwaverly · 2025-09-21T02:47:08Z

QA

Basic cycle tracing through logs

huntarr.sonarr - INFO - === Starting SONARR cycle ===
huntarr - INFO - Started cycle for sonarr (cyclelock = True)
huntarr.sonarr - INFO - Processing sonarr instance: sonarr
huntarr.sonarr - DEBUG - State check for sonarr.sonarr: 2.0 hours since last reset (interval: 24h)

no additional errors ✅
reset interval picked up correctly ✅

huntarr.sonarr - INFO - === SONARR cycle finished. No items processed in any instance. ===
huntarr.sonarr - INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
huntarr.sonarr - INFO -   sonarr: 12 items tracked, next reset: 2025-09-21 19:12:33 (24h interval)

error fixed ✅
accurate summary with reset interval being picked up correctly ✅

Interacting with state management through UI

no regressions in fetching summary ✅
no regressions in updating mode or reset interval ✅
no regressions in resetting state ✅

After setting a 1hr interval and resetting state:

sqlite> select * from stateful_instance_locks;
id|app_type|instance_name|created_at|expires_at|expiration_hours|updated_at
51|sonarr|sonarr|1758418052|1758421652|1|2025-09-21 01:27:32

correct reset time ✅
correct expiry ✅
correct expiration hours ✅

sqlite> select * from app_configs where app_type = "sonarr";
id|app_type|config_data|created_at|updated_at
77|sonarr|{
  "instances": [
    {
      "name": "sonarr",
      ...
      "state_management_mode": "custom",
      "state_management_hours": 1,
    }
  ],
  ...
}|2025-09-21 01:27:27|2025-09-21 01:27:27

correct state management hours ✅

Automatic state reset with new settings.

cycle 1:

huntarr.sonarr - INFO - === Starting SONARR cycle ===
huntarr.sonarr - INFO - Processing sonarr instance: sonarr
huntarr.sonarr - DEBUG - State check for sonarr.sonarr: 0.1 hours since last reset (interval: 1h)
...
huntarr.sonarr - INFO - === SONARR cycle finished. Processed items across instances. ===
huntarr.sonarr - INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
huntarr.sonarr - INFO -   sonarr: 1 items tracked, next reset: 2025-09-20 22:27:32 (1h interval)

cycle 2:

huntarr.sonarr - INFO - === Starting SONARR cycle ===
huntarr.sonarr - INFO - Processing sonarr instance: sonarr
huntarr.sonarr - DEBUG - State check for sonarr.sonarr: 0.3 hours since last reset (interval: 1h)
...
huntarr.sonarr - INFO - === SONARR cycle finished. Processed items across instances. ===
huntarr.sonarr - INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
huntarr.sonarr - INFO -   sonarr: 2 items tracked, next reset: 2025-09-20 22:27:32 (1h interval)

cycle 3:

huntarr.sonarr - INFO - === Starting SONARR cycle ===
huntarr.sonarr - INFO - Processing sonarr instance: sonarr
huntarr.sonarr - DEBUG - State check for sonarr.sonarr: 0.5 hours since last reset (interval: 1h)
...
huntarr.sonarr - INFO - === SONARR cycle finished. Processed items across instances. ===
huntarr.sonarr - INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
huntarr.sonarr - INFO -   sonarr: 3 items tracked, next reset: 2025-09-20 22:27:32 (1h interval)

cycle 4:

huntarr.sonarr - INFO - === Starting SONARR cycle ===
huntarr.sonarr - INFO - Processing sonarr instance: sonarr
huntarr.sonarr - DEBUG - State check for sonarr.sonarr: 0.7 hours since last reset (interval: 1h)
...
huntarr.sonarr - INFO - === SONARR cycle finished. Processed items across instances. ===
huntarr.sonarr - INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
huntarr.sonarr - INFO -   sonarr: 4 items tracked, next reset: 2025-09-20 22:27:32 (1h interval)

cycle 5:

huntarr.sonarr - INFO - === Starting SONARR cycle ===
huntarr.sonarr - INFO - Processing sonarr instance: sonarr
huntarr.sonarr - DEBUG - State check for sonarr.sonarr: 0.8 hours since last reset (interval: 1h)
...
huntarr.sonarr - INFO - === SONARR cycle finished. Processed items across instances. ===
huntarr.sonarr - INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
huntarr.sonarr - INFO -   sonarr: 5 items tracked, next reset: 2025-09-20 22:27:32 (1h interval)

cycle 6:

huntarr.sonarr - INFO - === Starting SONARR cycle ===
huntarr.sonarr - INFO - Processing sonarr instance: sonarr
huntarr.sonarr - DEBUG - State check for sonarr.sonarr: 1.0 hours since last reset (interval: 1h)
...
huntarr.sonarr - INFO - === SONARR cycle finished. Processed items across instances. ===
huntarr.sonarr - INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
huntarr.sonarr - INFO -   sonarr: 1 items tracked, next reset: 2025-09-20 23:27:32 (1h interval)

state reset ✅

castielwaverly · 2025-09-22T03:29:39Z

Well, that was apparently a bunch of code that didn't do anything.

castielwaverly · 2025-09-22T03:31:10Z

frontend/static/css/new-style.css

    padding: 12px 10px !important;
 }

-/* Make number inputs for intervals shorter */


None of these classes exist in the codebase anymore

castielwaverly · 2025-09-22T03:32:10Z

frontend/static/css/style.css

    transform: translateX(20px); /* Changed to match login page toggle (20px) */
 }

-/* Stateful Management Styling */


Same. Classes aren't used in the code anymore.

castielwaverly · 2025-09-22T03:37:17Z

frontend/static/js/new-main.js

-        // Stateful management reset button
-        const resetStatefulBtn = document.getElementById('reset_stateful_btn');
-        if (resetStatefulBtn) {
-            resetStatefulBtn.addEventListener('click', () => this.handleStatefulReset());
-        }
-
-        // Stateful management hours input
-        const statefulHoursInput = document.getElementById('stateful_management_hours');
-        if (statefulHoursInput) {
-            statefulHoursInput.addEventListener('change', () => {
-                this.updateStatefulExpirationOnUI();
-            });
-        }


Elements don't exist anymore -- dead code.

castielwaverly · 2025-09-22T03:38:48Z

frontend/static/js/new-main.js

-                if (settings.stateful_management_hours && document.getElementById('stateful_management_hours')) {
-                    const intervalInput = document.getElementById('stateful_management_hours');
-                    const intervalDaysSpan = document.getElementById('stateful_management_days');
-                    const expiresDateEl = document.getElementById('stateful_expires_date');
-
-                    // Update the input value
-                    intervalInput.value = settings.stateful_management_hours;
-
-                    // Update the days display
-                    if (intervalDaysSpan) {
-                        const days = (settings.stateful_management_hours / 24).toFixed(1);
-                        intervalDaysSpan.textContent = `${days} days`;
-                    }
-
-                    // Show updating indicator
-                    if (expiresDateEl) {
-                        expiresDateEl.textContent = 'Updating...';
-                    }
-
-                    // Also directly update the stateful expiration on the server and update UI
-                    this.updateStatefulExpirationOnUI();


Element doesn't exist anymore -- dead code.

castielwaverly · 2025-09-22T03:40:41Z

frontend/static/js/new-main.js

@@ -3477,227 +3427,30 @@ let huntarrUI = {
        // or use the stored configuredApps status if checkAppConnection updates it.
        this.checkAppConnections(); // Re-check all connections after a save might be simplest
    },
-
-    // Load stateful management info


None of the components on the receiving end of this data exist anymore. Code technically alive, but doing work for no reason.

castielwaverly · 2025-09-22T03:41:20Z

frontend/static/js/new-main.js

        console.log(`[getUserTimezone] Using timezone: ${timezone}`);
        return timezone;
    },
-


You can trigger any of these anymore -- dead code.

castielwaverly · 2025-09-22T03:43:34Z

src/primary/apps/eros/missing.py

 Exclusively supports the v3 API.
 """

-import time


Three types of changes:

Removing legacy state management

Removing unused variables and imports

Refactor to lazy formatting for logs

castielwaverly · 2025-09-22T03:43:56Z

src/primary/apps/eros/upgrade.py

 Exclusively supports the v3 API.
 """

-import time


Three types of changes:

Removing legacy state management

Removing unused variables and imports

Refactor to lazy formatting for logs

castielwaverly · 2025-09-22T03:44:43Z

src/primary/apps/eros_routes.py

@@ -1,43 +1,39 @@
 #!/usr/bin/env python3

-from flask import Blueprint, request, jsonify


Three types of changes:

Removing legacy state management

Removing unused variables and imports

Refactor to lazy formatting for logs

castielwaverly · 2025-09-22T03:46:16Z

src/primary/apps/lidarr/missing.py

 Handles missing albums or artists based on configuration.
 """

-import time


Three types of changes:

Removing legacy state management

Removing unused variables and imports

Refactor to lazy formatting for logs

castielwaverly · 2025-09-22T03:47:01Z

src/primary/apps/lidarr/upgrade.py


 import time
 import random
-from typing import Dict, Any, Optional, Callable, List, Union, Set # Added List, Union and Set


Three types of changes:

Removing legacy state management

Removing unused variables and imports

Refactor to lazy formatting for logs

castielwaverly · 2025-09-22T03:48:10Z

src/primary/apps/lidarr_routes.py

@@ -1,19 +1,17 @@
 #!/usr/bin/env python3

+import socket


Two types of changes:

Removing unused variables and imports

Refactor to lazy formatting for logs

castielwaverly · 2025-09-22T03:49:46Z

src/primary/apps/radarr/__init__.py

 # Module exports
 from src.primary.apps.radarr.missing import process_missing_movies
 from src.primary.apps.radarr.upgrade import process_cutoff_upgrades
-


Just a quick drive-by cleanup on this one.

castielwaverly · 2025-09-22T03:50:25Z

src/primary/apps/radarr_routes.py

@@ -1,19 +1,17 @@
 #!/usr/bin/env python3

+import socket


Two types of changes:

Removing unused variables and imports

Refactor to lazy formatting for logs

castielwaverly · 2025-09-22T17:56:56Z

src/primary/notification_manager.py

 """

 import logging
-import json


Just a quick drive-by cleanup in this file.

castielwaverly · 2025-09-22T17:57:50Z

src/primary/scheduler_engine.py

-def _get_user_timezone():
-    """Get the user's selected timezone from general settings"""
-    try:
-        from src.primary.utils.timezone_utils import get_user_timezone
-        return get_user_timezone()
-    except Exception:
-        import pytz
-        return pytz.UTC


Using get_user_timezone directly

castielwaverly · 2025-09-22T17:59:17Z

src/primary/background.py

+            # --- State Reset Check --- #
+            if has_instance_state_expired(app_type, instance_name):
+                app_logger.info("State has expired for %s instance '%s'. Resetting state.", app_type, instance_name)
+                reset_instance_state_management(app_type, instance_name)


Doing an intentional per-instance reset now.

castielwaverly · 2025-09-22T18:00:15Z

src/primary/scheduler_engine.py

-            # Check for stateful management expiration first
-            # Stateful management check debug removed to reduce log spam
-            check_stateful_expiration() # Call the imported function
-


No more global reset.

castielwaverly · 2025-09-22T18:04:29Z

src/primary/settings_manager.py

+def load_instance_settings(app_name: str, instance_name: str) -> dict[str, Any]:
+    """
+    Load settings that apply to a specific instance of an app.
+
+    Args:
+        app_name: The app name (sonarr, radarr, etc.)
+        instance_name: The specific instance name to load settings for
+
+    Returns:
+        dict: Dictionary containing the instance settings
+    """
+    app_settings = load_settings(app_name)
+
+    for instance_settings in app_settings.get("instances", []):
+        if instance_settings.get("name") == instance_name:
+            return instance_settings
+
+    raise ValueError(f"Instance '{instance_name}' not found for app '{app_name}'")


This is done all over the place, so might as well formalize it.

castielwaverly · 2025-09-22T18:05:18Z

src/primary/settings_manager.py

-        return False
+            settings_logger.info("Timezone applied to %s (TZ environment variable)", timezone)

-def validate_timezone(timezone_str: str) -> bool:


This is implemented in timezone_utils, so using that one instead.

castielwaverly · 2025-09-22T18:06:02Z

src/primary/state.py

@@ -1,234 +0,0 @@
-#!/usr/bin/env python3


👋 Bye, legacy² system

castielwaverly · 2025-09-22T18:06:29Z

src/primary/stateful_manager.py

-from src.primary.utils.database import get_database
-from src.primary.settings_manager import get_advanced_setting
-
-def initialize_lock_file() -> None:


👋 Bye, legacy system

castielwaverly · 2025-09-22T18:07:22Z

src/primary/stateful_manager.py

-
-def get_state_management_summary(app_type: str, instance_name: str, instance_hours: int = None) -> Dict[str, Any]:
+
+def get_instance_state_management_summary(app_type: str, instance_name: str) -> dict[str, Any]:


This now figures everything out itself, instead of requiring the caller for part of the summary.

castielwaverly · 2025-09-22T18:08:30Z

src/primary/stateful_manager.py

+

-def get_next_reset_time_for_instance(instance_hours: int, app_type: str = None) -> Optional[str]:
+def reset_instance_state_management(app_type: str, instance_name: str) -> bool:


New business logic for resetting instance state here.

castielwaverly · 2025-09-22T18:09:37Z

src/primary/stateful_routes.py

-# Create blueprint
 stateful_api = Blueprint('stateful_api', __name__)

-@stateful_api.route('/info', methods=['GET'])


This was called, but the info was never used. The client code calling this was removed above.

castielwaverly · 2025-09-22T18:10:23Z

src/primary/stateful_routes.py

-        response.headers['Access-Control-Allow-Origin'] = '*'
-        return response
+
+def base_response(status: int, data: dict[str, Any]):


This was done all over the palce in this file. Might as well define what a response should be in one place,

castielwaverly · 2025-09-22T18:10:49Z

src/primary/stateful_routes.py

+    return response
+

 @stateful_api.route('/reset', methods=['POST'])


This only does per-instance resets now.

castielwaverly · 2025-09-22T18:11:25Z

src/primary/stateful_routes.py

-        response.headers['Access-Control-Allow-Origin'] = '*'
-        return response
-
-@stateful_api.route('/update-expiration', methods=['POST'])


Nothing called this anymore.

castielwaverly · 2025-09-22T18:12:04Z

src/primary/utils/app_utils.py

@@ -1,42 +0,0 @@
-import socket


That file wasn't referenced anymore.

castielwaverly · 2025-09-22T18:13:26Z

src/primary/utils/database.py

            else:
-                logger.warning(f"Schedule {schedule_id} not found for update")
-
-    # State Management Methods


Legacy² system functionality removed.

castielwaverly · 2025-09-22T18:14:50Z

src/primary/utils/database.py

-    global _database_instance
-    if _database_instance is None:
-        _database_instance = HuntarrDatabase()
-    return _database_instance
+    if not hasattr(get_database, "db_instance"):
+        get_database.db_instance = HuntarrDatabase()
+    return get_database.db_instance


Global was only used in one function, so no point of having a global variable for this. Function attributes should work better here.

castielwaverly · 2025-09-22T18:15:18Z

src/primary/utils/logger.py

-    def _get_user_timezone(self):
-        """Get the user's selected timezone from general settings"""
-        try:
-            from src.primary.utils.timezone_utils import get_user_timezone
-            return get_user_timezone()
-        except Exception:
-            # Final fallback if timezone_utils can't be imported
-            import pytz
-            return pytz.UTC


Using get_user_timezone directly now

castielwaverly · 2025-09-22T18:15:37Z

src/primary/utils/timezone_utils.py

        return tz
-
-
-def get_timezone_name() -> str:


castielwaverly · 2025-09-22T18:16:37Z

src/primary/web_server.py

    """Redirect to main index with user section"""
    return redirect('./#user')
-
-# This section previously contained code for redirecting paths to include the base URL


All of this is dead code that is hidden after the return you see right above this section.

castielwaverly · 2025-09-22T18:17:14Z

src/primary/web_server.py

-        # Update expiration timing from general settings if applicable
-        try:
-            new_hours = int(data.get('stateful_management_hours'))
-            if new_hours > 0:
-                general_logger.info(f"Updating stateful expiration to {new_hours} hours.")
-                update_lock_expiration(hours=new_hours)
-        except (ValueError, TypeError, KeyError):
-            # Don't update if the value wasn't provided or is invalid
-            pass
-        except Exception as e:
-            general_logger.error(f"Error updating expiration timing: {e}")


Remove legacy system update.

castielwaverly · 2025-09-23T02:59:56Z

main.py

+        # Initialize state management system
+        try:
+            from src.primary.stateful_manager import initialize_state_management
+            initialize_state_management()
+        except Exception as e:
+            huntarr_logger.warning("Failed to initialize state management system: %s", e)


Just initialize once on startup, instead of trying to do it on every action

castielwaverly · 2025-09-23T03:01:17Z

src/primary/web_server.py

+        # Initialize state management for any new instances configured
+        initialize_state_management()


Also initialize on any updates. This and the startup initialization should cover existing + new tables that could be missing.

castielwaverly

More testing

Fresh install

First boot up

Logs:

INFO:stateful_manager:Initialized state management for sonarr/Default with 168h interval
ERROR:stateful_manager:Skipping initialization of instance 0 in radarr app. Missing setting: 'enabled'
INFO:stateful_manager:Initialized state management for lidarr/Default with 168h interval
INFO:stateful_manager:Initialized state management for readarr/Default with 168h interval
INFO:stateful_manager:Initialized state management for whisparr/Default with 168h interval
INFO:stateful_manager:Initialized state management for eros/Default with 168h interval

DB:

sqlite> select * from stateful_instance_locks;
1|sonarr|Default|1758643604|1759248404|168|2025-09-23 16:06:44
2|lidarr|Default|1758643605|1759248405|168|2025-09-23 16:06:45
3|readarr|Default|1758643605|1759248405|168|2025-09-23 16:06:45
4|whisparr|Default|1758643605|1759248405|168|2025-09-23 16:06:45
5|eros|Default|1758643605|1759248405|168|2025-09-23 16:06:45

Looks good. Note the radarr error is actually an error with the default settings missing the "enabled" key. In general, it's good to report unexpected things, instead of making a best guess an swallowing the error, hence the choice to avoid defaulting to False in this case.

Saving a new instance

Logs:

INFO:stateful_manager:Initialized state management for sonarr/new instance test with 48h interval

DB:

sqlite> select * from stateful_instance_locks;
1|sonarr|Default|1758643604|1759248404|168|2025-09-23 16:06:44
2|lidarr|Default|1758643605|1759248405|168|2025-09-23 16:06:45
3|readarr|Default|1758643605|1759248405|168|2025-09-23 16:06:45
4|whisparr|Default|1758643605|1759248405|168|2025-09-23 16:06:45
5|eros|Default|1758643605|1759248405|168|2025-09-23 16:06:45
6|sonarr|new instance test|1758643766|1758816566|48|2025-09-23 16:09:26

Looks good ✅. New instance is there with correct interval.

Running a cycle

Logs:

INFO - === [SONARR] Thread starting ===
INFO - === Starting SONARR cycle ===

INFO:stateful_manager:State check for sonarr.new instance test: 0.0 hours since last reset (interval: 48h)

INFO:stateful_manager:is_processed check: sonarr/new instance test, ID:58_1, Found:False, Total IDs:0
INFO:stateful_manager:is_processed check: sonarr/new instance test, ID:56_9, Found:False, Total IDs:0
INFO:stateful_manager:is_processed check: sonarr/new instance test, ID:57_1, Found:False, Total IDs:0
INFO - Found 3 unprocessed seasons out of 3 total seasons with cutoff unmet episodes.

INFO:stateful_manager:is_processed check: sonarr/new instance test, ID:58_1, Found:False, Total IDs:0
INFO:stateful_manager:[add_processed_id] Added ID 58_1 to database for sonarr/new instance test
INFO:stateful_manager:is_processed check: sonarr/new instance test, ID:4323, Found:False, Total IDs:1
INFO:stateful_manager:[add_processed_id] Added ID 4323 to database for sonarr/new instance test
INFO:stateful_manager:is_processed check: sonarr/new instance test, ID:4320, Found:False, Total IDs:2
INFO:stateful_manager:[add_processed_id] Added ID 4320 to database for sonarr/new instance test
INFO:stateful_manager:is_processed check: sonarr/new instance test, ID:4318, Found:False, Total IDs:3
INFO:stateful_manager:[add_processed_id] Added ID 4318 to database for sonarr/new instance test

INFO - === SONARR cycle finished. Processed items across instances. ===
INFO - === STATE MANAGEMENT SUMMARY FOR SONARR ===
INFO -   new instance test: 4 items tracked, next reset: 2025-09-25 17:06:39 (48h interval)
INFO - RESULT: Items processed successfully. Total tracked across instances: 4

DB:

sqlite> select * from stateful_processed_ids;
1|sonarr|new instance test|58_1|2025-09-23 17:07:11
2|sonarr|new instance test|4323|2025-09-23 17:07:11
3|sonarr|new instance test|4320|2025-09-23 17:07:11
4|sonarr|new instance test|4318|2025-09-23 17:07:11

Looks good ✅.

castielwaverly · 2025-09-23T18:04:39Z

And for a final sanity check: deploying these changes to my original, untouched 8.2.10 application has no unintended side effects. Both legacy tables are removed successfully, nothing gets reinitialized erroneously, state management is still tracking what it was already tracking, and state older than a week that's on a longer interval is maintained.

I'll be running this version on my system instead and fix anything that comes up. Otherwise, gonna call it here.

castielwaverly added 9 commits September 19, 2025 17:18

feat: Add instance-level state management

f56c1d9

refactor: Remove dead code, lazy format logs, align style with PEP8

dd403ca

refactor: Remove dead code, lazy format logs, align style with PEP8

83fbe67

refactor: Remove dead code, align style with PEP8

df2a85a

fix: Simplify and fix state management summaries

b729afe

refactor: Clean up stateful routes, align style with PEP8

37b5a50

wip: Use correct key

787355c

wip: Integrate flow better into existing functionality

671dcee

refactor: Use lazy formatting, improve formatting, log from app

a8abe4a

castielwaverly changed the title ~~feat: Add instance-level state management~~ fix: Fix instance level state management Sep 21, 2025

castielwaverly marked this pull request as ready for review September 21, 2025 02:51

castielwaverly marked this pull request as draft September 22, 2025 02:26

refactor: 🧹

54feb02

castielwaverly changed the title ~~fix: Fix instance level state management~~ refactor: Clean up state management Sep 22, 2025