enable using sccache
Another thing that might help container builds alongside #277 (closed) would be to enable sccache. Right now you need to hit the same runner that was used last time to get a good ccache, and as we increase the runner count that happens less often. If we had a storage bucket on packet that we could point sccache to, it could mean that once we've compiled the specific job once, we effectively don't have to ever again.