The easiest way to over-engineer a personal site is to treat it like a company platform. Notes can become a CMS migration. A blog can become a publishing pipeline. A simple deploy can become a release process with more ritual than risk.
I'm trying to keep the opposite rule in place: use the smallest tool that preserves the promise. If the promise is weekly notes, the tool should make writing feel like editing one object and pushing one commit. Anything heavier has to earn its keep.
This is product judgment wearing an engineering jacket. Good systems are not the ones with the most impressive machinery. They are the ones whose machinery matches the cost of being wrong. Add weight when the system earns it, not when your imagination gets bored.