[halbritt]

"On the cloud providers it is fully managed so nothing really to troubleshoot"

Clearly, you haven't run much at scale or experienced many problems.

Troubleshoot a network issue through an ingress, kube-proxy and then down to the pod.

Troubleshoot an issue with kube-api or etcd when you have no access to either the master or the etcd cluster.

Troubleshoot a load issue on a node related to IO.