-
-
Notifications
You must be signed in to change notification settings - Fork 291
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fix: gossipsub to yield more to the macro queue #5664
fix: gossipsub to yield more to the macro queue #5664
Conversation
Performance Report✔️ no performance regression detected Full benchmark results
|
attaching the profile of network thread of this branch network_thread_2023-06-19T08:39:44.576Z.cpuprofile.zip there are still some |
* fix: add setTimeout to onGossipsubMessage and onValidationResult * chore: more comments
🎉 This PR is included in v1.9.1 🎉 |
On the topic of I think this is due to the fact that we yield to the event loop after I/O polling most of the time so it does not really make a difference. In fact, I think There might be some potential if we only use cc @matthewkeil |
🎉 This PR is included in v1.10.0 🎉 |
Motivation
Description
part of #5556
part of #5604
Test result on mainnet node
the 1st event loop lag metric is not a big difference , but it's still better than unstable (from 12h)
the 2nd event loop lag is clearly better
especially the incoming request handler time
also outgoing request time
it helps our peers more stable, more peers want to stay with us