Worth linking to the upstream tickets related to this:
+1, from my side. What are the real hard technical requirements? Can we run of something outside of Fedora/CentOS DC? Do we really need openshift/kube cluster?....
The ideal IMO would be an OpenShift cluster with multi-arch nodes. It's unlikely we're going to get that at the beginning though (see https://github.com/coreos/fedora-coreos-tracker/issues/262#issuecomment-5519...).
So to start with... just really any way we can schedule work onto multi-arch nodes. There's a gradient there of better to less ideal options. The fedora-infrastructure ticket for example is about leveraging Koji. But really even if it's just SSH, we could work with that.