As a result of the ongoing Jenkins outages / PSI resource problems... I've started to rethink how we can do multiple steps within the same node.

So rather than having jobs call other jobs, eg.

• build assets in one job*,
  • fetch assets and build in Brew in a second job, then
    • copy to quay in a third job

... we'd combined these operations into a single sync-and-build job, such that we result in "holding" the nodes longer than today, but also avoid spinning up 3 nodes for one build (assets, image, push to quay).

* - Exceptions to this may include jobs which require multi-arch assets, like configbump, traefik, and theia, as we need a way to build / fetch those assets on the correct arch, then use them in Brew. Migrating everything into the same get-sources script breaks that ability, unless we can run the asset build/fetch step in one stage, then when all 3 arches are done, commit locally updated asset files ONCE, and only run the brew build ONCE for all 3 updated arches.

For the multiarch asset cases, we might only succeed in having a successful build include a push to quay rather than triggering a downstream job.