Our self-hosted runner is set up on Kubernetes and uses this image. We've noticed that running multiple execution breaks because the runner doesn't seem to do a full cleanup in between jobs. The docs specify that github hosted runners "provide a clean instance for every job execution" but how would we go about that for self-hosted runners? Is there an automated way to do this or would we have to manually clean things up before each run?
For self-hosted runner, currently have to do the cleanup manually for each run, thanks.