Dedicated infrastructure for building triggered packages
As part of https://gitlab.com/gitlab-org/omnibus-gitlab/issues/2322 .
Create a dedicated runner manager for package, docker image builds, and QA testing.
It would be great if we could get the k8s runner manager running so we can do some more efficient resource usage. Also, this would help build caching which would ultimately speed up the builds.
This is a prerequisite for enabling package builds by default for all CE/EE MRs.
I think this is a nice way for more cross-team collaborations. We could use some help from CI team on configuring k8s runner and some help from the infrastructure team on getting the infrastructure created. My estimate is that a very little time from each dedicated team member would be required, we just need to have a point of contact we can reach.
Having the k8s runner in a (currently) non-essential service would help us get a nice testing ground before we fully move or specs and package builds to k8s.
@pcarranza @ayufan what do you think about this? Would it be possible to get someone from your team to help out as a side task?