Hi all, since a few days we have issues with our Rancher hosted on a 3-node k3s cluster. Not really sure but it might be related to some performance issue with etcd - I think we never called compact or defrag on the etcd. is this necessary? if so, do i need to keep some old revisions?