> and sometimes the fix is not an as easy as restarting a server.
Yes, but more often than not you just restart a couple of nodes, maybe kill and re up a couple containers and that's about it. The question is do the cases where you really have to dig in outweigh those where restarts just work?