Hi! Hetty author here. I was just experimenting with SQLite as we speak! While I like Cayley so far, I'm already hitting some performance issues when doing full text search/queries that need regex matching on large datasets. Given that HTTP request logging is perfectly suitable for a non-graph database, SQLite is probably more pragmatic. Also, I'm hoping that using a SQLite3 extension for regex (C) will be more performant than using Go's `regexp` package. And of course there's the added benefit you mentioned of a widely used DB with lots of (external) tooling.

The only downsides I see so far are that the project will need cgo, and thus cross compilation is a bit trickier for doing releases on GitHub. But already found a possible tool for making it easier: https://github.com/troian/golang-cross

This would be fun at work to find most hit uri, slowest url and the like.

Logging on the a back end is fine but this would be simpler I think.