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
Make it easy to kubectl describe the workload that evicted another workload.
Why is this needed:
Kueue Workloads have events like kueue-admission Preempted to accommodate a workload (UID: <UID>) due to fair sharing within the cohort
As far as I can tell kubectl --selector field only supports filtering on values in the object .metadata.labels. The Workload UID is in the .metadata.UID field which can not be filtered on by kubectl --selector. For clusters with few workloads grepping on kubectl get workloads -A would work but does not scale for clusters with thousands of workloads.
Completion requirements:
Make it easy to kubectl describe the Workload that preempted another workload. Simple solution: add the Workload UID to the Workload labels.
This enhancement requires the following artifacts:
Design doc
API change
[ x] Docs update -> add docs on how to get details of preempting workload.
The artifacts should be linked in subsequent comments.
The text was updated successfully, but these errors were encountered:
I don't remember if kueuectl would allow it, probably not. However, in any case as an external binary the only way to find the preempting workload by UID would be to list all workloads first, which is not scalable indeed.
I like the idea of adding the UID as a label. cc @tenzen-y WDYT?
What would you like to be added:
Make it easy to
kubectl describe
the workload that evicted another workload.Why is this needed:
Kueue Workloads have events like
kueue-admission Preempted to accommodate a workload (UID: <UID>) due to fair sharing within the cohort
As far as I can tell kubectl --selector field only supports filtering on values in the object .metadata.labels. The Workload UID is in the .metadata.UID field which can not be filtered on by kubectl --selector. For clusters with few workloads grepping on kubectl get workloads -A would work but does not scale for clusters with thousands of workloads.
Completion requirements:
Make it easy to
kubectl describe
the Workload that preempted another workload. Simple solution: add the Workload UID to the Workload labels.This enhancement requires the following artifacts:
The artifacts should be linked in subsequent comments.
The text was updated successfully, but these errors were encountered: