0 downtime. Jesus Christ. Nginx and HAProxy solved this shit decades ago. You can drop out a server or group. Deploy it. Add it back in. With a single telnet command. You don’t need junk containers to solve things like “0 downtime deployments”. That was a solved problem.

Calm down my friend!

You are not wrong, but that only covers a part of what I was asking. How about the rest? How do you actually bring your services to production? I'm curious.

And, PS, I don't use k8s. Just saying.