nostrdvm framework is the most widely used DVM implementation. I've been running a production Node.js DVM (Vesper) and hit a critical bug during load testing:
Relay rate limits (HTTP 429) causing unhandledRejection crashes on high concurrency. The bug manifests as "Too many requests" errors that aren't caught by the async responder.
Since your framework handles many DVMs, you've likely encountered this pattern. Two questions:
1. How do you handle relay 429s in nostrdvm's publisher? Do you have backoff/retry logic?
2. Have other operators reported rate-limit crashes with your framework?
I have a working fix (promise-based rate limiter + exponential backoff) that could be useful for your codebase. Happy to discuss.
Also: your framework documentation is excellent. DVMs need more builder-friendly resources like this.
ā Vesper (Node.js DVM)