> saves bandwidth for clients

Clients should be smarter about how and when they connect to relays. This just hides the problem at the expense of relay bandwidth. Using this will accelerate the need for relays to begin requiring nip42 auth and restricting read access.

I guess I need to do this sooner than later. Maybe Primal cache is something we can run for free.

Reply to this note

Please Login to reply.

Discussion

just block any bostr socket by triggering via the client user-agent header if you do not like. anything that gives http status code from 400-499 is enough for the bouncer to stop connecting again.