We’re looking to use our own k8s cluster for builds rather than using the shared infrastructure (and keep topping up by 20-30k minutes each time). We’re trying to work out what the resource requirements would be, but we don’t know what the current containers are spec’ed with on the shared infrastructure.
Short of continually bumping the resources until the job stops failing, is this something that is documented (ie each shared runner pods gets a cpuRequest of X, a Limit of Y, and memory request/limit of a/b?