Dev Ojha

That makes sense! I'm down for deleting per peer id metrics as well

This sounds great to me, I really like it! Can you add it to these two spots: - https://github.com/cometbft/cometbft/blob/d27732ad9a992e321e440417a2a8220f90fe0044/internal/consensus/reactor.go#L1444-L1449 - https://github.com/cometbft/cometbft/blob/main/p2p/conn/connection.go#L361-L362 Then can help with getting a mainnet profile for...

Lets measure after that gets in! The memory overhead is coming from the prometheus `With` and `Add` calls I believe, so the creation of the PeerId as a string isn't...

Commented out the spot instances in Osmosis fork for noticeable performance gains / reductions in gossip routines and consensus mutex!

I think we should re-open this issue. I think from the issue we were leaning towards just removing these peer metrics entirely. Or should I make a new issue for...

Happy to try PR'ing a change if this sounds acceptable to folks. (Having a single clock variable, and a background process started on `init` for updating it. Alternatively could make...

No we should re-open this, that PR only closed the first checkbox of this issue!

Oh we have a really easy fix for ` Proto writer is supposed to use an internal buffer. We make a new object and then must re-allocate the internal buffer...

> How would you uniquely identify a BlockID? By the Hash (which end-ups being a []byte)? That seems ideal to me, as long as the block hash also commits the...

Turns out I was wrong about the batch verification vs not status! (However I am still correct about the triple verification) I got this wrong from: (1) Using v0.34.x benchmarks,...