My builds are now at a stage where most of the time is spent downloading the docker container a workflow/job should be run on.
Attempts to mitigate this includes using “standard” building blocks (alpine) and open sourcing my build/test containers to speed things up. I still have a 90% host cache miss rate. I can only assume you have a pretty high cache eviction rate in your cluster, but its still unfortunate to see things like this:
…do you have suggestions or best practices to increase bootup times?