Migrate away from gcr.io/k8s-testimages repo

[ameukam]

gcr.io/k8s-testimages is currently a Google-owned GCR/AR repo. We should migrate to a community-owned AR registry.

/priority important-longterm

[BenTheElder]

x-ref kubernetes/k8s.io#1523

Images in gcr.io/k8s-testimages are not being patched, and at some point in the near future it will be shut down.

Please move to current public images, see for example kubernetes/k8s.io#1523 (comment)

[mkumatag]

currently these are the dependencies:
https://github.com/search?q=repo%3Akubernetes%2Fperf-tests%20gcr.io%2Fk8s-testimages&type=code

• gcr.io/k8s-testimages/probes
• gcr.io/k8s-testimages/netperfbenchmark
• gcr.io/k8s-testimages/perf-tests-util/access-tokens
• gcr.io/k8s-testimages/perf-tests-util/containerd
• gcr.io/k8s-testimages/perf-tests-util/request-benchmark
• gcr.io/k8s-testimages/quay.io/prometheus-operator/prometheus-config-reloader
• gcr.io/k8s-testimages/quay.io/prometheus/node-exporter
• gcr.io/k8s-testimages/grafana/grafana

[mkumatag]

Most of these images come from https://github.com/kubernetes/perf-tests/tree/master/util-images. A few observations:

• I don’t see any post-submit jobs to build and push these images when changes happen. Should we consider using Cloud Build to trigger builds via CI?
• Are all these images still required, or can some of them be replaced with others?
• Do we need to create a new staging repository for these images, or can we use an existing one, such as gcr.io/k8s-staging-e2e-test-images?

[ameukam]

I don’t see any post-submit jobs to build and push these images when changes happen. Should we consider using Cloud Build to trigger builds via CI?

I would say yes. That should help with low-key maintenance of those images.

Are all these images still required, or can some of them be replaced with others?

I don't have the expertise to properly answer this but I suspect we still all the images.

Do we need to create a new staging repository for these images, or can we use an existing one, such as gcr.io/k8s-staging-e2e-test-images?

I think a new staging repo should be fine.

[mkumatag]

I would say yes. That should help with low-key maintenance of those images

I just realised that we have an existing staging repository: gcr.io/k8s-staging-perf-tests. We simply need to set up the Cloud Build mechanism to build and push these images to that location; there's no need to create a new one.

[ameukam]

/reopen

[k8s-ci-robot]

@ameukam: Reopened this issue.

In response to this:

/reopen

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-sigs/prow repository.

[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