I'm not telling people to avoid running a LLM at home. I'm just saying that it wouldn't be a generic purpose one close to what ChatGPT provides. The reason I would guess is primarily a lack of computational power.

Some generic purpose LLM probably.

I wanted to learn how things work.

First of all you need to establish an SLA. As for automation you have all kinds of things to choose from, like puppet, ansible, terraform, salt, pulumi, argocd, flux, etc (not necessarily exclusive). Personally I like terraform to provision resources and Ansible to configure them. Some I do through Gitea actions.

Sounds to me like you're asking for a michelin rated meal at McDonalds.

https is not a problem. But you'll need an internal CA and distributed its certificate to your hosts' trust store.