This message was deleted.

My first question would be etcd. If that's 5 nodes then I'd look at how to shrink it to 3 first. I don't think the control plane cares too much past that, but definitely etcd gets cranky if it has fewer nodes than it expects.

Try: Rancher Manager -> Cluster Management -> Clusters. This assumes that this question is in regards to a cluster deployed by Rancher.

yea this is a k3s cluster, i never tried the scale down before, will have to take a look at that

Yes, this assumes a Rancher provisioned, downstream cluster.

If it's k3s and isn't using embedded etcd, you can potentially just delete the spare control plane nodes one by one and it should balance.