What you describe is from the "pets" era of server deployment, and we are now deep into the "cattle" era. Train yourself on destroying and redeploying, and building observability into the stack from the outset, rather than managing a server through ssh. Every shop you go to professionally is going to work like this. Eventually, Linux desktops will work like this also, especially with all the work going into systemd to support movable home directories, immutable OS images with modular updates, and so forth.