Hey guys hope you’re well. I’ve been reading and testing quicknode performance for mempool txns. Have some questions if someone from the team can take a look:
- Quicknode says that they can’t pull the entire p2p network for txpool. Makes sense it would be difficult. However, I do see that quicknode has some of the smallest discovery, yet is the fastest. For reference I’ve written my tests in js and python too see if there’s any difference. (very little 3 ms, I would call it library overhead).
Are there optimal ways to pull from the pending tx pool? Quick alerts seems to only run inspect on blocks that are already confirmed. Pool inspect and content are extremely compute heavy. Streaming new members to the pool is always ideal but if we’re only getting 10% of what’s out there it’s rubbish. I
Note: I’ve implemented your guide as a benchmark, with the same performance.