David Barsky

Note from backlog bonaza: This feature is perma-unstable. The Lang Team is not in any rush to remove this feature, but if a new stabilization-track feature appears that overlaps with...

> @davidpdrsn tower::buffer::Buffer is supposed to propagate spans https://docs.rs/tower/latest/src/tower/buffer/service.rs.html#148-151, so I'd be surprised if that's the issue `tower::buffer::Buffer` is propagating a `tracing::Span`, but @arttet is creating an [OpenTelemetry span](https://github.com/arttet/simple-service/blob/11fdc71795c894e6eab331519c3ca05b60f09d02/simple-service/src/service/simple_service.rs#L32-L33). @arttet:...

I haven't reviewed this yet, but can you make this PR target the master branch? We generally handle backports; it's easier for us if this targets master.

> On master the interests cache doesn't exist, so this is not an issue there? Ugh, that's an oversight. That _should_ be on master.

mercurial + the diff tool encourages the equivalent of force pushes, but github punishes you for it, so no worries.

This is a pretty helpful change, thanks. Upon working more with `tracing-glog`, would it be a good idea to expose some of the simpler bits of configuration (e.g., thread names,...

This is somewhat resolved by #2110—the core issue is that the root span in the example is closed _after_ the collector shuts down due to drop rules.

I wouldn't rule out a breaking change, @CAD97—if the performance improvement is as good as I think that it might be, then I think we should seriously considering release a...

I think there's a few other items that merit a breaking release: https://github.com/tokio-rs/tracing/milestone/11

Oh, sorry! I meant that the breakage would be in `tracing-subscriber`, which can maybe be used as a dry-run-of-sorts for cross-version compatibility of Layers before employing that in tracing 0.2.