Discrepancy in tracked peers

[ladidan]

Description

I've recently noticed a significant difference in logged peer count per INFO logs (e.g. [...] Peers: 40) vs. various prometheus/grafana metrics-dashboards (querying the API) which after a certain amount of peer discovery time settle at around the max-peer defined number (e.g. 25).

Versions (Add all that apply)

22.7.0-RC3 on Ubuntu 22.04 server

[iamhsk]

Hi @ladidan, thank you for opening an issue. Could you answer the following questions to help us investigate this further?

• Which network were you running besu on?
• If running private network, were you using boot nodes?
• Could you provide us a log extract?

[ladidan]

• tested on Goerli
• log would look like e.g.: EthScheduler-Workers-3 | INFO | PersistBlockTask | Imported #7,339,454 / 26 tx / 0 om / 10,402,057 (34.7%) gas / (0x1afc02c542fc345d3286121f7ce25d16103262c3dcfcc832eed303b3a65b780b) in 0.232s. **Peers: 121**

whereas the grafana dashboard shows 25 current peers (via the besu API)

[iamhsk]

Hi @ladidan, there is already an open issue to resolve this. Please follow #2689 for progress :)