my goal is to have a fault tolerant and distributed relay

this uses a new architecture that decouples the relay instances from the database, and some optimizations that can reduce load and space usage a lot

It also allows the easy addition of indexes for anything like search, and allows subdividing a relay into millions of partitions that can be accessed individually or in groups nostr:note125d6wlcq55cvz8sg4xjkhsp5e76lqzypcuyzytl3808u9t0f0jkqpjwmgk