go-bitswap
go-bitswap copied to clipboard
Improve bitswap debug logging
In an effort to support diagnosing issues with Bitswap, particularly in deployed nodes, we should add more debug level logging. So that it's possible to get more post-mortem information on what happened when IPFS nodes become slow.
Does debug logging really solve the issue of post-issue diagnosis on deployed nodes? That would require having logging always on debug level (at least for bitswap), which will have its own associated overhead?