This introduces a "soft_timeout" config option, which specifies after which ammount of seconds updatebot should abort its run. The check for this happens after each task. In some cases this can prevent the external taskcluster timeout from killing updatebot while it is working on a task, resulting in a unexpected databse state.
This would also require a follow-up patch on mozilla-central for adding a config value for the soft_timeout here and potentially updating the "hard" taskcluster timeout here. I could imagine setting the soft timeout to 60 minutes and the hard timeout to 90 minutes (currently the hard timeout is 60 minutes).
This introduces a "soft_timeout" config option, which specifies after which ammount of seconds updatebot should abort its run. The check for this happens after each task. In some cases this can prevent the external taskcluster timeout from killing updatebot while it is working on a task, resulting in a unexpected databse state.
This would also require a follow-up patch on mozilla-central for adding a config value for the soft_timeout here and potentially updating the "hard" taskcluster timeout here. I could imagine setting the soft timeout to 60 minutes and the hard timeout to 90 minutes (currently the hard timeout is 60 minutes).