You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, if a runner disappears in the middle of a job, that job will just hang indefinitely.
The broker should notice when a runner isn't communicating. It could definitely amend the log, and it perhaps should also mark the job as unclaimed so that another runner can try.
The text was updated successfully, but these errors were encountered:
This is still a real issue. We should probably add a timeout in the broker to mark a claimed task as unclaimed if it sits idle for more than a couple of minutes.
Now, if there's just something really wrong with a fiddle or with a runner, we could get into an endless loop of [ runner claims -> runner crashes -> runner is restarted -> broker unclaims ]. We may want to consider a circuit breaker of some sort to ensure this pathological case doesn't run more than a few loops.
Currently, if a runner disappears in the middle of a job, that job will just hang indefinitely.
The broker should notice when a runner isn't communicating. It could definitely amend the log, and it perhaps should also mark the job as unclaimed so that another runner can try.
The text was updated successfully, but these errors were encountered: