Add NodeClass label to nodes for scheduling and observability

[jonathan-innis]

Description

What problem are you trying to solve?

There's currently an issue open in aws/karpenter-provider-aws that's requesting the addition of the karpenter.k8s.aws/ec2nodeclass as a well-known label that can be used for scheduling and assigned to nodes for observabillity: https://github.com/aws/karpenter-provider-aws/issues/5178. We think that this could be generalized across all cloud providers.

We currently don't enforce the kind or apiVersion of the nodeClassRef but as we are progressing to v1, I think that we should consider requiring users specify this. If we enforced that users had to specify this field, we could add this label for scheduling and node label propagation as <group>/<kind>: <name>

How important is this feature to you?

This improves observability for users as called out in: https://github.com/aws/karpenter-provider-aws/issues/5178

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

[github-actions[bot]]

This issue has been inactive for 14 days. StaleBot will close this stale issue after 14 more days of inactivity.

[sftim]

How about making it an annotation, not a label? That way, the annotation could store a JSON blob that uniquely identifies the group, kind and name of the node class.

If we want a simple string, I would use the URL path of that resource in Kubernetes' API. eg /apis/karpenter.k8s.aws/*/ec2nodeclasses/example or a string with an obvious derivation such as karpenter.k8s.aws/ec2nodeclasses/example.

Alternatively, use two labels, eg: karpenter.sh/node-class-api-path: /apis/karpenter.k8s.aws/v1beta1/ec2nodeclasses karpenter.sh/node-class-name: example

(which would still allow for compound selectors)

[k8s-triage-robot]

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

This bot triages un-triaged 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:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue 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.

This bot triages un-triaged 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:

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

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

/lifecycle rotten

[jonathan-innis]

/remove-lifecycle rotten

[k8s-triage-robot]

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

This bot triages un-triaged 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:

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

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

/lifecycle stale