Enable large github runners for etcd-io org

[jmhbnz]

Organization or Repo

etcd-io/etcd

User affected

@serathius, @ahrtr

Describe the issue

After migrating the etcd-io org out from under the CNCF enterprise account to under the Kubernetes enterprise account, large runners need to be re-enabled: https://docs.github.com/en/actions/using-github-hosted-runners/about-larger-runners/about-larger-runners

etcd ci workflows are currently unable to complete, refer: https://github.com/etcd-io/etcd/actions/runs/6915180325/job/18813834949?pr=16959

etcd requires large runners ubuntu-latest-8-cores for robustness testing.

CNCF projects have accesses to larger instances, but it has to be explicitly enabled by an admin of https://github.com/etcd-io :

• Open https://github.com/organizations/etcd-io/settings/actions/runner-groups-
• Click Default Larger Runners
• Click Allow public repositories
• Click Selected repositories, and choose the etcd & bbolt repositories

[ahrtr]

• Click Selected repositories, and choose the etcd repo

bbolt also needs larger runner.

[mrbobbytables]

I had to create the large runner definition^^;; I'm guessing those were created previously in the CNCF enterprise account. Let me know if you have any issues.

[jmhbnz]

I had to create the large runner definition^^;; I'm guessing those were created previously in the CNCF enterprise account. Let me know if you have any issues.

Confirming the larger runners are now working once more. Thanks @mrbobbytables 🙏🏻

[ahrtr]

Thanks @mrbobbytables for the quick fix.

[ahrtr]

I see that one bbolt workflow (runs on ubuntu-latest-8-cores) is getting stucked for almost 2 hours https://github.com/etcd-io/bbolt/actions/runs/6928339142/job/18843944830

@jmhbnz @mrbobbytables

[serathius]

Observed same on the robustness tests.

[mrbobbytables]

sorry about all this back and forth =/
I know the issue now - I'm going to delete the runners scoped in the etcd org and create them at the k8s enterprise org and still grant access to the same repos. It's where we can easily centrally manage the limits and such.

[mrbobbytables]

Okay, I've updated it - I've set the billing limit at 2k for now. Let me know if you have problems. We MAY need an azure subscription for it to all work correctly >_> I've pinged a few folk but with the US holiday things might not get looked at till next week. The GitHub admin team also has a meeting with GitHub next Tuesday so any issues we encounter we can surface there as an escalation.

/reopen
(till we know everything is good)

[k8s-ci-robot]

@mrbobbytables: Reopened this issue.

In response to this:

Okay, I've updated it - I've set the billing limit at 2k for now. Let me know if you have problems. We MAY need an azure subscription for it to all work correctly >_> I've pinged a few folk but with the US holiday things might not get looked at till next week. The GitHub admin team also has a meeting with GitHub next Tuesday so any issues we encounter we can surface there as an escalation.

/reopen
(till we know everything is good)

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.

[ahrtr]

Thanks @mrbobbytables for the quick fix. Confirmed that the bbolt workflow is running well right now.

[jmhbnz]

Large ubuntu-latest-8-cores runners have been working nicely since 21st November 🙏🏻