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
The KEDA autoscaling feature is quite powerful for high concurrent workloads when using a celery executor.
However, KEDA supports defining multiple triggers to base your desired replicas for.
I believe there are some cases in which the desired replicas might be higher than the task amount might indicate.
For this one could leverage for example the memory utilization metric of the airflow-worker (if you have specified appropriate resource requests).
As an example you can see a triggers: section for the ScaledObject here :
This would require an extension of the existing ScaledObject
I'm not sure if the keda autoscaling feature is widely used and whether the effort to include this would be justified.
Note that as a bypass you could patch the ScaledObject after the helm deployment is finished to include the additional metrics.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
The KEDA autoscaling feature is quite powerful for high concurrent workloads when using a celery executor.
However, KEDA supports defining multiple triggers to base your desired replicas for.
I believe there are some cases in which the desired replicas might be higher than the task amount might indicate.
For this one could leverage for example the memory utilization metric of the airflow-worker (if you have specified appropriate resource requests).
As an example you can see a triggers: section for the ScaledObject here :
This would require an extension of the existing ScaledObject
I'm not sure if the keda autoscaling feature is widely used and whether the effort to include this would be justified.
Note that as a bypass you could patch the ScaledObject after the helm deployment is finished to include the additional metrics.
Beta Was this translation helpful? Give feedback.
All reactions