Repo-proxy pods have restarted way to often on k8s which is not a normal behavior.
As we have defined a LivenessProbe and a ReadynessProbe in the deployment resource, it means that the HTTP check must be improved
k8s@jenkins-radish:~$ .bin/kubectl get pods -l app=repo-proxy --kubeconfig .kube/config
NAME READY STATUS RESTARTS AGE
repo-proxy-1417053348-k40ie 1/1 Running 144 92d
repo-proxy-1417053348-rk73p 0/1 Running 55 92d