fix(workers): worker-level recovery — autoretry on transient errors + tightened sweep (5min) + import_media_file body-wrap so no path leaves rows stuck in 'processing'

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

backend/app/tasks/maintenance.py

@@ -15,7 +15,7 @@ from ._sync_engine import sync_session_factory as _sync_session_factory
 
 log = logging.getLogger(__name__)
 
-STUCK_THRESHOLD_MINUTES = 30
+STUCK_THRESHOLD_MINUTES = 5
 OLD_TASK_DAYS = 7
 PHASH_PAGE = 500
 VERIFY_PAGE = 200
@@ -24,11 +24,15 @@ FFPROBE_TIMEOUT_SECONDS = 10
 
 @celery.task(name="backend.app.tasks.maintenance.recover_interrupted_tasks")
 def recover_interrupted_tasks() -> int:
-    """Find ImportTask rows stuck in 'processing' for >30 min and re-queue them.
+    """Find ImportTask rows stuck in 'processing' for >5 min and re-queue them.
 
-    Why 30 min: large videos can legitimately take many minutes to import;
-    30 is a safe gate that catches actual crashes (which leave the row stuck
-    forever) without resetting slow-but-still-running jobs.
+    Why 5 min: import_media_file is sub-second for the vast majority of
+    files; even a large-video transcode caps at the per-task soft_time_limit
+    (5 min) defined on the task itself. Anything still 'processing' after
+    that window is a confirmed crash (worker died, DB disconnect mid-flush,
+    OOM) and must be recycled. Was 30 min historically; tightened
+    2026-05-24 after operator hit a 2224-row zombie pile during the IR
+    migration scan.
     """
     SessionLocal = _sync_session_factory()
     cutoff = datetime.now(UTC) - timedelta(minutes=STUCK_THRESHOLD_MINUTES)