Helm Chart template and Image CI for Karpenter Provider CAPI

[lukren]

It would be nice to have a Helm Chart template for karpenter-provider-cluster-api like the provider for aws: https://github.com/aws/karpenter-provider-aws/tree/main/charts

This would first require establishing an image CI that would automate the building and pushing of image versions for the Karpenter Cluster API provider controller to an image registry.

[elmiko]

thank you for proposing this @lukren , i agree. having a helm chart would be a nice option for users, i think we need to get an automated build for our images setup so that we can use a trusted registry in the helm charts.

but, with that said, we could certainly start to write some charts and leave the image spec as a template option. i'm not a helm expert, but we should be able to do something like that.

[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

[elmiko]

/remove-lifecycle rotten

[elmiko]

given that #38 is addressing the helm chart, i would like to discuss with the cluster-api community about hosting the image (thank you @ttsuuubasa for the suggestion). if they cannot host the image, we might be able to use quay.io as the cluster-api kubemark provider uses that as well.

essentially, could we use registry.k8s.io for karpenter-clusterapi-controller?

[ttsuuubasa]

During the August 27, 2025 meeting (recording), the answer was “yes.”

We need to follow the documentation: registry.k8s.io documentation

Initially, I thought that we would need to request the creation of a staging repository for karpenter-provider-cluster-api.
Is my understanding correct?

Or should we instead reuse the existing staging repository for cluster-api?

[elmiko]

Initially, I thought that we would need to request the creation of a staging repository for karpenter-provider-cluster-api.
Is my understanding correct?

i believe you are correct. looking at the directory for k8s-staging-cluster-api, it appears to be only cluster-api controllers. i think good practice for us would be to create a new directory k8s-staging-karpenter-provider-cluster-api.

[ttsuuubasa]

Requesting to reopen this issue because the image CI task has not yet been completed.

To request the creation of the staging repository k8s-staging-karpenter-provider-cluster-api, we need to update the following files in the kubernetes/k8s.io repository:

• infra/gcp/infra.yaml
  Add k8s-staging-karpenter-provider-cluster-api to the staging.projects list.

• groups/sig-cluster-lifecycle/groups.yaml
  Create a Google Group for access to the staging repository:
  k8s-infra-staging-karpenter-provider-cluster-api@kubernetes.io
  Although the registry.k8s.io documentation only mentions editing infra.yaml (which is sufficient for auto-push via Google Cloud Build), having manual access via the group is more convenient and maintainable.
  Just to confirm, karpenter-provider-cluster-api belongs to sig-cluster-lifecycle, correct?

• groups/restrictions.yaml
  Add the above Google Group email-id to define access restrictions.

Would it be possible for @elmiko to create this PR?
Given your standing in the community, I believe the request would carry more weight coming from you than from me.