At the moment raising an exception is the only way to fail a job. I'm open to ideas.

I don't know much about the implementation of rq. As a user, I imagine something like:

from rq import get_current_job

def mytask():
    job = get_current_job()
    job.fail("That’s why I fail.")

I don't know if I should open a new issue, but I think it's strongly related to this one:

Is there a possibility to raise an exception from "outside" of the worker or job in order to cancel/stop an ongoing job?

Is there a possibility to raise an exception from "outside" of the worker or job in order to cancel/stop an ongoing job?

This issue is about failing a job from “inside” the job. So I think the question of how to stop it from the outside merits its own issue. But perhaps you can reference this one.

Please see this answer.

Also, I use a custom retry handler function implementation. The job of the retry handler is,

• Gracefully handle the custom Exception (in my case it was ThirdPartyException class)
• If there is a status.HTTP_429_TOO_MANY_REQUESTS, re-enqueue the failed job after the API Rate limit duration.

I think its worth having a look and think it might help in some way.

https://github.com/premkumar30/vendor-retailer/blob/b3d806569dc46e489bbf497f995dfc9492cfd492/vendor_retailer/utils.py#L44

Starting from version 1.5.0, RQ allows you to configure retries.

queue.enqueue(my_func, retry=Retry(max=3, interval=60))

So if you want RQ to retry your function, simply raise an exception within your function.

I think there's no other solution than using exceptions because running code after failing the job doesn't make sense.

I can implement a function similar to #1313 (comment) that will raise an error with some more information than a simple error.

So the good news is that the in development version of RQ actually solves a few issues discussed in this thread.

This issue is about failing a job from “inside” the job. So I think the question of how to stop it from the outside merits its own issue. But perhaps you can reference this one.

RQ 1.7.0 comes with send_stop_job_command that allows you to stop a currently executing job from the outside. https://python-rq.org/docs/jobs/#stopping-a-currently-executing-job

Also, I use a custom retry handler function implementation

RQ 1.8.0 (to be released) will include on_success and on_failure callbacks so you can decide what to do yourself whenever a job finishes or fails. The PR is https://github.com/rq/rq/pull/1480/files