Miretpl commented on PR #69807: URL: https://github.com/apache/airflow/pull/69807#issuecomment-4974094621
> It is. The Github issue says I was on my phone and didn't take a deeper look at the issue. Thanks! > I think it waited for helm experts to share their thoughts if this is something we want or not. I tend to say no but @jedcunningham @Miretpl I guess it's up to you As of now, I would say no as a result backend (Airflow database), and the queue are separated in the Helm implementation (and probably in most Airflow deployments, I would say). Even if there is no queue instance, tasks may continue proper execution without interruption if the result backend is working correctly and if Celery handles the restarted queue; everything should be good. If we kill the Celery pod via probes when the queue has temporary issues, that will mean killing tasks during execution too, which, depending on the case, might be a disruptive operation which should not happen during temporary issues. I will take a deeper look at the issue itself later this week. There is always a chance that I miss something 😄 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
