-
Notifications
You must be signed in to change notification settings - Fork 260
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support for "safe-to-evict" annotation #984
Comments
Another use case I came across is long-running Kubernetes Jobs and ASG instance refresh. It would be great to wait for job to finish here. |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If you want this issue to never become stale, please ask a maintainer to apply the "stalebot-ignore" label. |
This issue was closed because it has become stale with no activity. |
Describe the feature
Kubernetes autoscalers use a "safe-to-evict" annotation (https://kubernetes.io/docs/reference/labels-annotations-taints/#cluster-autoscaler-kubernetes-io-safe-to-evict) to identify pods that should not be evicted while they are running. Would it be possible to add support for this, or a similar solution, in NTH? In many cases it would probably need a grace period after which a forceful drain would occur (e.g. in case of spot disruption).
Is the feature request related to a problem?
We have Kubernetes Jobs that are running tests. Sometimes these jobs are terminated e.g. by rebalance recommendations, that could wait a short while for the job to finish. This disrupts our CD workflows.
Describe alternatives you've considered
Outside of being open to other labels/annotations being used instead of the default "cluster-autoscaler" one, I do not have alternative ideas
The text was updated successfully, but these errors were encountered: