That is one problem, yes. A related problem is it’s entirely too easy to set up a K8s cluster and/or add extras, while having little to no understanding of how

That is one problem, yes. A related problem is it’s entirely too easy to set up a K8s cluster and/or add extras, while having little to no understanding of how the technology functions. It’s difficult to troubleshoot K8s when you have no idea what cgroups are, or how iptables works, etc. The problem gets worse with managed K8s, because now the control plane is removed from your list of concerns, but it can still cause problems. So now you can still get errors from etcd doing a compaction / defragmentation run, but the logs telling you that are buried, and you probably don’t know what etcd is anyway.