Remove unneeded node groups regardless of scale down being in cooldown. #5901
Conversation
k8s-ci-robot
added
the
kind/feature
k8s-ci-robot
added
area/cluster-autoscaler
size/S
|
Unknown CLA label state. Rechecking for CLA labels. Send feedback to sig-contributor-experience at kubernetes/community. /check-cla |
k8s-ci-robot
added
the
cncf-cla: yes
| if err != nil { | ||
| klog.Errorf("Error while removing unneeded node groups: %v", err) | ||
| } | ||
| scaleDownStatus.RemovedNodeGroups = removedNodeGroups |
There was a problem hiding this comment.
I think this makes line 653 no-op now.
There was a problem hiding this comment.
I don't think so, as in line 648 we are overwriting scaleDownStatus
There was a problem hiding this comment.
Ah, good point. This function is way too long 😞
|
/lgtm |
k8s-ci-robot
added
the
lgtm
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: olagacek, x13n The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
k8s-ci-robot
added
the
approved
…of scale down being in cooldown into CA1.26
…of scale down being in cooldown into CA1.27
…am-cluster-autoscaler-release-1.26 Backport #5901 [CA] Remove unneeded node groups regardless of scale down being in cooldown into CA1.26
…am-cluster-autoscaler-release-1.27 Backport #5901 [CA] Remove unneeded node groups regardless of scale down being in cooldown into CA1.27
What type of PR is this?
/kind feature
What this PR does / why we need it:
Currently unneeded node groups are deleted only if scale down isn't in cooldown, so in particular if there was a scale up in a given loop, no unneeded node group will get deleted. There are certain scenarios in which that might turn out problematic - eg if scale up fails in every loop, the node groups won't be deleted for a long time. At the same time, unneeded node groups might affect scale up failures (e.g. quota issues).
Due to that, we can decouple unneeded node group deletion from the rest of scale down logic.
Does this PR introduce a user-facing change?
Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.: