Sorry I wasn’t clear.

By ´As a dev´ I meant that listing three files was not going to help you when things went wrong in production, and you had bizarre network problems because of problems with your cni, kernel vs kubernetes issues, etc. I understand fully it can be made to work, but not by everyone and not everyone has the time to do so.

So I take from your explanations that as a k8s user I have two options : have very good infra engineers, or rent k8s from a cloud provider.

( side note : I work primarily in hpc, and have found that having strong infra engineers makes a world of difference , and many software engineers don’t appreciate it enough )