k8s-infra-prow-build-trusted jobs should require requests/limits be set

[BenTheElder]

We're clearly not enforcing this currently, unlike the other k8s-infra clusters.

It will be a pain to do retroactively, but we really should be doing this, so we can effectively autoscale the cluster.

/sig k8s-infra testing

[marcelgaedke]

Hi @BenTheElder , I am new here and have been looking at this repo for a while and would like to start contributing. This isn't labeled as good first issue, but it seems doable. I would start by parsing all files under "config/jobs" to find all jobs running on "k8s-infra-prow-build-trusted" which are missing requests/limits. I may need some advice on how to determine proper values for requests/limits for each job. And I think the actual validation/enforcement for new jobs happens around "config/tests/jobs" ? If it's okay I would like to start working on this.

[BenTheElder]

I may need some advice on how to determine proper values for requests/limits for each job.

This is the problem indeed, we already have code to enforce this, but we can't turn it on for this cluster without figuring out that part, and access to where these jobs run is highly sensitive (it runs secure workloads for automating the project), I don't think we have a good way to grant access to just see the usage for these currently.

[BenTheElder]

This is why I didn't label it, because I don't know how to make that part accessible, and the folks with access are pretty focused on other work at the moment 😅

(in particular: https://groups.google.com/a/kubernetes.io/g/dev/c/p6PAML90ZOU)

[BenTheElder]

We also don't have agreement from any other leads yet, this issue hasn't been discussed yet. Normally we'd require agreement before requesting help. I appreciate the offer to help, but this probably isn't the best issue.

[marcelgaedke]

Alright. 👍 Thanks for taking the time to explain. I'll keep my eyes open for good first issues.

[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