fix: Do not process job; instead re enqueue if shutdown was requested

[vikigenius]

Fixes #280

[tobymao]

it's unclear exactly what problem this is solving. can you write a test?

[vikigenius]

The idea is that if you were waiting a while for dequeue, a signal could have arrived in the meantime, but if you go ahead and start the job without checking you just started a job even though shutdown was requested.

I will take a look at the tests and write one to showcase the behavior.

[tobymao]

i don't think you should re-enque it, it should probably just fail and let the retry handle it

[vikigenius]

You are right, I don't think it's good to enqueue again and just introduces potential complexity. If we are letting it fail, we don't need to do anything, it will auto fail after timeout anyway I will close the PR