fix: guard against multiple cleanup threads and register AppConfig

- Add default_app_config in __init__.py so Django picks up the AppConfig
- Use a module-level lock to ensure only one cleanup thread starts
  (gunicorn calls ready() multiple times for master + workers)
- Remove debug print statements

Author: Boanerges1996

app/__init__.py

@@ -0,0 +1 @@
+default_app_config = 'app.apps.UsersConfig'

app/apps.py

@@ -7,24 +7,33 @@
 
 logger = logging.getLogger(__name__)
 
+_cleanup_started = False
+_cleanup_lock = threading.Lock()
+
 
 class UsersConfig(AppConfig):
-    name = 'api'
+    name = 'app'
 
     def ready(self):
-        interval = int(os.getenv('CONFERENCE_CLEANUP_INTERVAL_SECONDS', 3600))
+        global _cleanup_started
 
+        interval = int(os.getenv('CONFERENCE_CLEANUP_INTERVAL_SECONDS', 3600))
         if interval <= 0:
             return
 
+        with _cleanup_lock:
+            if _cleanup_started:
+                return
+            _cleanup_started = True
+
         def cleanup_loop():
-            # Wait before first run to let the app fully start
             time.sleep(60)
 
             while True:
                 try:
                     from app.management.commands.cleanup_stale_conferences import Command
-                    Command().handle(hours=int(os.getenv('CONFERENCE_TIMEOUT_HOURS', 4)), dry_run=False, verbosity=1)
+                    hours = int(os.getenv('CONFERENCE_TIMEOUT_HOURS', 4))
+                    Command().handle(hours=hours, dry_run=False, verbosity=0)
                 except Exception as e:
                     logger.warning(f'[ConferenceCleanup] Error: {e}')