From MediumCo I will say: there is a lot of friction, quite intentionally, when you go off the rails. As a backend engineer, messing with things like VMs and SSL certificates yourself is definitely way off the rails. We want that to be hard. You are supposed to be creating and iterating on services using the standard application frameworks, deployed to the standard shared clusters, communicating through the standard service mesh, etc. where all of that is handled automatically.

Make sure there's not a golden path that you're missing? Or that your reasons for departing from it are really insurmountable?