fix: improve worker task failure handling

Use asyncio.wait with FIRST_EXCEPTION to:
- Properly name tasks for better error logging
- Cancel remaining tasks when one fails
- Log which specific manager failed before propagating the exception

backend/worker.py

@@ -1,6 +1,7 @@
 """Background job worker using pgqueuer."""
 
 import asyncio
+import contextlib
 import json
 import logging
 import random
@@ -169,11 +170,27 @@ async def main() -> None:
 
         logger.info("Worker started, processing queue jobs and scheduled jobs...")
 
-        # Run both managers concurrently
-        await asyncio.gather(
-            qm.run(),
-            sm.run(),
+        # Run both managers concurrently - if either fails, both stop
+        queue_task = asyncio.create_task(qm.run(), name="queue_manager")
+        scheduler_task = asyncio.create_task(sm.run(), name="scheduler_manager")
+
+        done, pending = await asyncio.wait(
+            [queue_task, scheduler_task],
+            return_when=asyncio.FIRST_EXCEPTION,
         )
+
+        # Cancel any pending tasks
+        for task in pending:
+            task.cancel()
+            with contextlib.suppress(asyncio.CancelledError):
+                await task
+
+        # Check for exceptions in completed tasks
+        for task in done:
+            exc = task.exception()
+            if exc is not None:
+                logger.error(f"Task '{task.get_name()}' failed: {exc}")
+                raise exc
     finally:
         await queue_conn.close()
         await scheduler_conn.close()