I managed a few production services like this and it added a lot of overhead to my work. On the one hand I'd get developers asking me why their stuff hasn't been deployed yet. But then I'd also have to think carefully about when to deploy and actually watch it to make sure it came back up again. I would often miss deployment windows because I was doing something else (my real job).

I'm sure there are many solutions but K8s gives us both fully declarative infrastructure configs and zero downtime deployment out of the box (well, assuming you set appropriate readiness probes etc)

So now I (a developer) don't have to worry about server restarts or anything for normal day to day work. We don't have a dedicated DevOps/platforms/SRE team or whatnot. Now if something needs attention, whatever it is, I put my k8s hat on and look at it. Previously it was like "hmm... how does this service deployment work again..?"