Conversation
wemeetagain
commented
Jan 11, 2024
- update libp2p to v1.x (and associated dependencies)
- our noise crypto optimizations have been upstreamed and are no longer needed
- discv5 is now split into two packages: enr and discv5 (for rationale see feat: create separate enr package discv5#267)
- revamp libp2p dashboard with relevant metrics
- deployed to feat3 on Jan 11
packages/beacon-node/package.json
Outdated
| "@chainsafe/libp2p-noise": "^13.0.1", | ||
| "@chainsafe/discv5": "^7.1.0", | ||
| "@chainsafe/enr": "^2.0.2", | ||
| "@chainsafe/libp2p-gossipsub": "^11.0.1", |
There was a problem hiding this comment.
@wemeetagain is it worth to use v11.1.0 for this upgrade? that's for batchPublish option ChainSafe/js-libp2p-gossipsub#482
There was a problem hiding this comment.
Yeah, we can upgrade it to 11.1.0. Should we turn on batchPublish in certain places?
There was a problem hiding this comment.
yes just need to add a new option to Eth2Gossipsub constructor batchPublish: true. That PR need to release first
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## unstable #6285 +/- ##
============================================
- Coverage 80.38% 80.38% -0.01%
============================================
Files 202 202
Lines 19620 19618 -2
Branches 1176 1176
============================================
- Hits 15771 15769 -2
Misses 3821 3821
Partials 28 28 |
Performance Report✔️ no performance regression detected Full benchmark results
|
|
I see some great metrics with this PR (mostly on the test mainnet node)
looks great overall 🚀 |
* wip * chore: update discv5 and enr * chore: bump dependencies * chore: fix import statement * chore: update libp2p dependencies * chore: yarn.lock cleanup * chore: update discv5 * feat: revamp libp2p dashboard * Clean up metrics registry type casts * chore: fix lint / build errors * chore: more import updates * chore: more yarn.lock cleanup * chore: update gossipsub --------- Co-authored-by: Nico Flaig <nflaig@protonmail.com>
|
🎉 This PR is included in v1.15.0 🎉 |
