What are (ballpark) rate limits on the public n0 relays?

[dpc]

I'm working on a little project and I'm testing it locally, and I ran two instances of a service on my dev machine so they talk with each other.

I've noticed the communication between them stops sometimes in a mysterious way. I started debugging my own code, but couldn't really find anything that could have caused it, and then I realized that if I slow down their activity with sleeps, everything seems to work reliably, so maybe I am hitting rate limits on the public gateways? Since they are on the same machine, they probably can't talk with each other directly(?), and I do let them run for some time when testing, so maybe they run out of quota.

I don't think they exchange a lot of data, but I have absolutely no idea what are the rate limits. How would I confirm my suspicions? And what should be the expectations of what probably will work OK using public relays?

Edit: Turns out the problem was actually keeping around a lot of protocol handlers that were waiting for an internal update, while the requesting party already disconnected, so they were accumulating and maybe hitting some limit.

[matheus23]

Since they are on the same machine, they probably can't talk with each other directly(?),

They should be able to! If these two processes can talk to each other via udp sockets, they should be able to talk to each other via iroh. If not, that's a bug.

(Sorry, I can't check the rate limits myself)