Clear stale sync_log 'running' entries on startup

After a container restart any in-progress jobs leave orphaned 'running' rows
with no completed_at. Added startup cleanup to mark them failed, so the UI
doesn't show a permanently stuck status.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/main.py

@@ -61,6 +61,29 @@ async def lifespan(app: FastAPI):
     except Exception as e:
         logging.getLogger(__name__).warning(f"Stale batch cleanup failed: {e}")
 
+    # Startup: mark any sync_log entries stuck as 'running' as failed
+    # (these are orphaned by container restarts mid-job)
+    try:
+        import psycopg2
+        from database import DATABASE_URL
+        conn = psycopg2.connect(DATABASE_URL)
+        try:
+            conn.autocommit = True
+            with conn.cursor() as cur:
+                cur.execute("""
+                    UPDATE sync_log
+                    SET status = 'failed',
+                        completed_at = NOW(),
+                        error_message = 'Interrupted by container restart'
+                    WHERE status = 'running' AND completed_at IS NULL
+                """)
+                if cur.rowcount:
+                    logging.getLogger(__name__).info(f"Cleared {cur.rowcount} stale sync_log entries")
+        finally:
+            conn.close()
+    except Exception as e:
+        logging.getLogger(__name__).warning(f"Stale sync_log cleanup failed: {e}")
+
     # Ensure ai_insights table exists
     try:
         from database import SyncSessionLocal