This is different from qu's currently released behavior which is to simply store failures for perusing at a later time. While most jobs can be retried and will succeed on a retry, there are some that would be retried forever. It would probably be good to allow the backend to decide what to do when a job raises an exception, rather than having the worker abort which will cause a retry.
This is different from qu's currently released behavior which is to simply store failures for perusing at a later time. While most jobs can be retried and will succeed on a retry, there are some that would be retried forever. It would probably be good to allow the backend to decide what to do when a job raises an exception, rather than having the worker abort which will cause a retry.