The default resource requests we set on application controller and application set workloads are too high for small clusters, due to the overhead OCP already puts on the nodes.

Currently, both do request 1 full CPU at scheduling time. The repo server requests 500m at scheduling time.

This should be lowered to a more reasonable value of 250m for all three, application controller, repo server and applicationset in order to allow scheduling of the workloads on small (possibly test or dev) clusters.