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
Is your feature request related to a problem? Please describe.
I use queue rate limiting to stay within limits of an external API I'm calling. Some jobs however do not require this call, eg the job might be too old and I just want to fail it. These jobs however, as far as I can see, count to the rate limit. So in the worst case a bunch of unimportant NOOP jobs that finish immediately prevent actual useful work from being done. In normal operation this would not be issue -- but in case there are some hickups that stop job processing for a longer duration having to process the stale jobs first would prolong the time to recovery.
Describe the solution you'd like
A way to return from the worker function that makes the job that was processed not count towards rate limits.
The text was updated successfully, but these errors were encountered:
hi @sakari, something was investigated some time ago #1653 and also we notice that it would affect rate limit. No sure when we would have time to revive it as it didn't get too much traction
Unfortunately this is a bit difficult to implement as the rate-limit is updated as soon as job is moved to active status. However we can keep this open as an enhancement as it is in the realms of possibility to implement it some day.
Is your feature request related to a problem? Please describe.
I use queue rate limiting to stay within limits of an external API I'm calling. Some jobs however do not require this call, eg the job might be too old and I just want to fail it. These jobs however, as far as I can see, count to the rate limit. So in the worst case a bunch of unimportant NOOP jobs that finish immediately prevent actual useful work from being done. In normal operation this would not be issue -- but in case there are some hickups that stop job processing for a longer duration having to process the stale jobs first would prolong the time to recovery.
Describe the solution you'd like
A way to return from the worker function that makes the job that was processed not count towards rate limits.
The text was updated successfully, but these errors were encountered: