Add test for retry exhaustion moving task to failed queue

[tejakusireddy]

Summary

Adds a non-live test covering retry exhaustion behavior for storage-backed queues.

The test verifies that when a task has no retries left and the callback fails, pull_and_execute() returns False, removes the task from the main queue, and moves the serialized task to the failed queue.

Testing

• ruff check tests/test_api.py
• pytest tests/test_api.py::test_retry_exhaustion_moves_task_to_failed_queue -q
• pytest tests/test_api.py -q
• git diff --check

[temporaer]

@tejakusireddy , if you plan to submit more regularly, installing the pre-commit hook might be worth it :-) it catches all these linting errors before committing/pushing.

[temporaer]

@tejakusireddy , can you describe why you think this particular test is important to have? Any ideas about making this work for service bus backend?

[tejakusireddy]

@temporaer Thanks, I’ll set up the pre-commit hook before sending more PRs.

For this test, I added it because retry exhaustion is the terminal failure path for storage-backed queues. When num_retries reaches 0, the worker should return False, remove the task from the active queue, and preserve the serialized task in the failed queue. I didn’t see a focused non-live test covering that behavior.

For Service Bus, the equivalent path would be DLQ/dead-letter behavior rather than the storage failed queue. I kept this PR storage-focused to avoid adding another live Service Bus test, but a follow-up could cover the Service Bus path either with a mocked backend/envelope unit test or a live DLQ test if that’s preferred.

[temporaer]

@tejakusireddy thank you. While adding tests is good work, especially bug fixes and feature contributions are welcomed.

[tejakusireddy]

@temporaer Makes sense, thanks for the guidance. I’ll focus future contributions more on small bug fixes or feature improvements, with tests included where appropriate.