will the kubernetes cluster work fine if the rancher server that manages it goes down ? i agree that there might be issue with scaling / auto healing of nodes , but the k3s shd work just fine with pod auto scaling enabled ryt ?