cluster-autoscaler: Add option to disable scale down for unready nodes

[gregth]

Which component are you using?:
Cluster Autoscaler

Is your feature request designed to solve a problem? If so describe the problem this feature should solve.:
There are cases where a user may not want the unready nodes to be removed from a cluster.

As an example, but not limited to, this might me useful in case a node is unreachable for a period of a time and local data live there, the node shall remain in the cluster, and possibly an admin may want to take any actions for recovering it.

Describe the solution you'd like.:
Provide the ability to disable scale-down for unready nodes. The Cluster Autoscaler ships with the --scale-down-unready-time flag, that configures "How long an unready node should be unneeded before it is eligible for scale down". Currently, if the flag is set to a negative or zero value, the unready nodes will be removed immediately. We could modify the logic of the code and if a negative value is provided, the cluster autoscaler shall disable scale down for unready nodes.

Describe any alternative solutions you've considered.:
Another alternative solution is to introduce a new flag for disabling scale down for unready nodes, but, we will end up with two flags that will be dependant.

Additional context.:

[gregth]

I have sent a PR to implement this feature, see #4877. Any feedback is more than welcome.

[drmorr0]

(imo) this is a breaking change that could cause somewhat surprising behaviour for cluster operators. If someone sets --scale-down-unready-time to a negative number now, and then they upgrade CA, this could cause them to never scale down nodes again, which is not desirable.

[gregth]

(imo) this is a breaking change that could cause somewhat surprising behaviour for cluster operators. If someone sets --scale-down-unready-time to a negative number now, and then they upgrade CA, this could cause them to never scale down nodes again, which is not desirable.

@drmorr0 Your concern is valid. But, as we have discussed in #4877 (comment), we opt to add a new separate flag for disabling the scale-down of unready nodes, instead of extending the functionality of the existing flag. We must -and will- keep this backwards compatible, the default behavior will be that the Cluster Autoscaler removes unready nodes, unless the user explicitly requests to not remove them via the new flag we'll introduce.

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue or PR as fresh with /remove-lifecycle stale
• Mark this issue or PR as rotten with /lifecycle rotten
• Close this issue or PR with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[k8s-triage-robot]

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue or PR as fresh with /remove-lifecycle rotten
• Close this issue or PR with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle rotten

[k8s-triage-robot]

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.

This bot triages issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Reopen this issue with /reopen
• Mark this issue as fresh with /remove-lifecycle rotten
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/close not-planned

[k8s-ci-robot]

@k8s-triage-robot: Closing this issue, marking it as "Not Planned".

In response to this:

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.

This bot triages issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Reopen this issue with /reopen
• Mark this issue as fresh with /remove-lifecycle rotten
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/close not-planned

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.