Interesting... if every user has their own database, then there are separate backups for each. Connection pooling isn't possible (although I guess one never use

Interesting... if every user has their own database, then there are separate backups for each. Connection pooling isn't possible (although I guess one never uses connection pooling with D1 anyway?) Schema changes need to be separately handled across each database. I guess there's a lot of considerations here. Cool stuff, although I would personally stick to a more "boring" architecture for my own work.