Clean up org teams and make them valuable for use around SIGs

[iholder101]

Our organization teams are dramatically outdated: https://github.com/orgs/kubevirt/teams.

They should be cleaned up, and be useful for pinging a sig, e.g. @sig-network-reviewers.

[iholder101]

/assign

FYI @aburdenthehand @dhiller

[dhiller]

Hey @iholder101, our GitHub teams are controlled by a GitOps by making changes to the orgs.yaml (located inside project-infra repository) and then opening a pull request.

In a nutshell a prow job running peribolos synchronizes changes to the orgs.yaml towards the GitHub orgs roughly every two hours.

You can find more information about our GitOps based org management here: https://github.com/kubevirt/community/blob/main/docs/automating-github-org-management.md

[dhiller]

I believe there's a duplicate issue that aims at a similar use case - will try to find it.

[dhiller]

It seems that idea was just in my head - related to this is kubevirt/project-infra#3814 which might adhere to the overall concept of automating sig sync and teams etc across repositories.

I'd imagine an automation that syncs GitHub teams with the OWNERS_ALIASES from k/kubevirt and creates approvers vs. reviewers teams as appropriate. This way we would spare the manual sync process. WDYT?

[dhiller]

Also to take into account should be the sigs.yaml from k/community which might be our single source of truth in the long term - I think we need to invest some thought into this.

[iholder101]

I'd imagine an automation that syncs GitHub teams with the OWNERS_ALIASES from k/kubevirt and creates approvers vs. reviewers teams as appropriate. This way we would spare the manual sync process. WDYT?

Sounds amazing! 👍

Also to take into account should be the sigs.yaml from k/community which might be our single source of truth in the long term - I think we need to invest some thought into this.

Yes 👍
The cross-org support is important IMO since many repos would be structured around sigs (k/k, k/community, k/enhancements, etc) so if we can have a single source of truth and sync it to other repos via automation I think it would be ideal. Indeed need to think about the specifics here.