Closed guseggert closed 2 years ago
Are test failures related to this change?
Yes, I need to fix them, meant to open this as a draft until I work those out
OK it's good to go now, the remaining test failures are known flaky tests documented in https://github.com/ipfs/go-libipfs/issues/86
LGTM
This adds tracing spans to the costly Bitswap entry points. It doesn't instrument the bitswap internals, which will take some time. In go-ipfs, this will at least let us know the contribution of Bitswap to the overall request handling time.
This also plumbs contexts through internally so that they reach the content routing APIs, so that traces are propagated through and we can start instrumenting e.g. the DHT.