Hello everyone! I have the following scenario: 1 cluster manager CAPI 1 downstream cluster provisioned by CAPI (not imported) It turns out that RANCHER was installed in the downstream cluster, which automatically applied several modifications, where RANCHER started to self-manage its own downstream cluster as management. The local namespace of rancher was removed and the cluster was broke, all control planes and worker nodes disappeared, however the VMs were not removed. They are invisible in "kubectl get nodes" and the pods are all in a pending state. The question is how can I install Rancher on a cluster provisioned by CAPI without it interfering with CAPI's management of the cluster. Has anyone ever experienced this?