This is a bootstrap issue to start the discussion about Sticky Runners and what we can do to drive it forward to get it delivered.
The idea of
Sticky Runners is to re-use existing machine when it is safe to run a build on it in order to increase the chance of having a docker image cache, job cache, artifacts. The end result is faster build times and reduced cost of running as we do follow up on the build.
Actual proposal will come later as we evaluate possible solutions and define an MVP.