Let me put this in perspective: the NYTimes bought Wordle for >$1m.
As for your specific case, you're talking about databases. The startup world is littered with the corpses of database startups so it's already a tough situation. On top of that you're talking about replication and query caching so of all the people who care deeply about databases (where those databases are generally free) you're talking about a small selection of them where something like this actually would matter.
Now you can make a business out of serving a small number of customers by providing something that is of incredible value to them but you will have to identify those customers and sell to them. Often such companies start because someone identifies such a problem, leaves one of those companies and starts a business to do just that. They already have the contacts to sell back to them.
The other side of this is that in the US as a senior engineer making $400k+ every year completely risk free with good work life balance and good benefits is relatiely easy.
You say you want to start a family. Assuming you're in the US then things like health insurance and paying for education are going to matter. But here's a big one: you will want to (or should want to) spend time with your children, not sweating over the possibility that you can't pay the mortgage this month.