A lot of the solutions to pod problems is to let kubernetes restart them. System administrators work to prevent problems.
Pods are also typically single purpose, rather than general purpose like a bare metal machine. System admin knowledge isn't required.
I don't think you need to unlearn a lot, but you do need to accept a different mindset, that of horizontal scaling and disposablility.
There's tons of people, me included, who run a server as a hobby. Almost nobody runs kubernetes as a hobby. It's like Hadoop - only corporations can love it.