How to increase timeout of failed jobs

[eriktelepovsky]

Hello. Is there a way how to change/increase timeout of failed jobs?

My jobs fail with exception:

rq.timeouts.JobTimeoutException: Task exceeded maximum timeout value

so I increased the DEFAULT_TIMEOUT in RQ_QUEUES setting, but every time I requeue failed jobs, the new setting is ignored and and is used previous timeout value.

[selwin]

Are you getting the queue using django_rq.get_queue()?

[eriktelepovsky]

Hi @selwin. Do you mean when I requeue the failed jobs? I don't do it programmatically. I use django admin action:

[selwin]

No, I meant when you're enqueueing the job originally.

[eriktelepovsky]

I'm using method decorator:

from django_rq import job

@job('long_running')
def my_method()
   ...

so yes, I presume it is using get_queue()

[selwin]

Yeah, so I think that should work. Are your settings.py configured correctly? Alternatively you can also do @job('long_running', timeout=1000)

[eriktelepovsky]

When I increase the timeout in the settings, it is applied to all new delayed jobs successfully. But it is not applied to the requeued failed jobs. Old timeout value is used.

[selwin]

Ah ok. Then what’s needed is a feature that allows you to update job.timeout prior to requeueing it.

Unfortunately this feature is not in DjangoRQ yet. If you want, I’ll be happy to accept a PR for this.

On Nov 30, 2020, at 3:54 PM, Erik Telepovský [email protected] wrote:

 When I increase the timeout in the settings, it is applied to all new delayed jobs successfully. But it is not applied to the requeued failed jobs. Old timeout value is used.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[eriktelepovsky]

I can look into it. Can you guide me where to start?

[selwin]

Create an edit_job_timeout view in https://github.com/rq/django-rq/blob/master/django_rq/views.py#L302

This view should be reachable via job detail page:

[eriktelepovsky]

And what about automatic timeout update according to the current settings when job is requeued?

[selwin]

It won't be automated, since the job object is already created and stored in Redis, complete with it's timeout. Automatically updating a value that's already stored, whether it's in DB or Redis whenever a configuration is changed is not recommended.

[eriktelepovsky]

I created a PR #458