update readmes for flame and error #695
Conversation
tracing-error/README.md
Outdated
| - [`InstrumentResult`] and [`InstrumentError`] extension traits, which | ||
| provide an [`in_current_span()`] method for bundling errors with a | ||
| [`SpanTrace`]. | ||
| - [`ExtractSpanTrace`] extension trait, for extracting `SpanTrace`s from | ||
| behind `dyn Error` trait objects. |
There was a problem hiding this comment.
I'm not sure if this needs to include all the traits enabled by the feature flag --- @davidbarsky what do you think?
Co-Authored-By: Eliza Weisman <eliza@buoyant.io>
There was a problem hiding this comment.
Looks good to me overall. I had a few last edits.
@davidbarsky, if you have a minute, would love your review on this as well.
| # tracing-flame | ||
|
|
||
| A Tracing [Layer][`FlameLayer`] for generating a folded stack trace for generating flamegraphs | ||
| and flamecharts with [`inferno`] |
There was a problem hiding this comment.
this isn't strictly inferno-specific; it emits the perf output format, right? so technically any flamegraph tool that expects that can use this crate's output.
i'm not sure if this is worth stating here though, for the sake of brevity...
There was a problem hiding this comment.
I don't think its the same format that perf emits, my understanding is that with perf normally you take the output and pass it through another script that folds the stack traces, and then you pass the folded representation into the flamegraph generating scripts. With tracing-flame we skip straight to emitting a folded representation. I think it might be cool to investigate if we can emit perf's exact format so we can be compatible with more tools, but I think it would be the same as writing a script to fake a bunch of stack traces that reproduce the folded repr, so for now I don't think we should bother with it.
There was a problem hiding this comment.
Yeah, this is the folded output produced by flamegraph, inferno, and some newer eBPF perf tracing tools. perf's output (by which you're probably referring to perf script's output) is probably not something we want to mirror. The collapsed output is decently standard, enough so that eBPF also produces the same thing.
Co-authored-by: Eliza Weisman <eliza@buoyant.io>
# 0.2.0 (October 23, 2021) This is a breaking change release in order to update the `tracing-subscriber` dependency version to [the v0.3.x release series][v03]. ### Changed - Updated `tracing-subscriber` dependency to [v0.3.0][v03] ([#1677]) ### Fixed - Disabled default features of the `tracing` dependency so that proc-macro dependencies are not enabled ([#1144]) - Documentation fixes and improvements ([#635], [#695]) ### Added - **SpanTrace**: Added `SpanTrace::new` constructor for constructing a `SpanTrace` from a `Span` passed as an argument (rather than capturing the current span) ([#1492]) Thanks to @CAD97 for contributing to this release! [v03]: https://github.com/tokio-rs/tracing/releases/tag/tracing-subscriber-0.3.0 [#635]: #635 [#695]: #695 [#1144]: #1144 [#1492]: #1492 [#1677]: #1677
# 0.2.0 (October 23, 2021) This is a breaking change release in order to update the `tracing-subscriber` dependency version to [the v0.3.x release series][v03]. ### Changed - Updated `tracing-subscriber` dependency to [v0.3.0][v03] ([#1677]) ### Fixed - Disabled default features of the `tracing` dependency so that proc-macro dependencies are not enabled ([#1144]) - Documentation fixes and improvements ([#635], [#695]) ### Added - **SpanTrace**: Added `SpanTrace::new` constructor for constructing a `SpanTrace` from a `Span` passed as an argument (rather than capturing the current span) ([#1492]) Thanks to @CAD97 for contributing to this release! [v03]: https://github.com/tokio-rs/tracing/releases/tag/tracing-subscriber-0.3.0 [#635]: #635 [#695]: #695 [#1144]: #1144 [#1492]: #1492 [#1677]: #1677
# 0.2.0 (October 23, 2021) This is a breaking change release in order to update the `tracing-subscriber` dependency version to [the v0.3.x release series][v03]. ### Changed - Updated `tracing-subscriber` dependency to [v0.3.0][v03] ([tokio-rs#1677]) ### Fixed - Disabled default features of the `tracing` dependency so that proc-macro dependencies are not enabled ([tokio-rs#1144]) - Documentation fixes and improvements ([tokio-rs#635], [tokio-rs#695]) ### Added - **SpanTrace**: Added `SpanTrace::new` constructor for constructing a `SpanTrace` from a `Span` passed as an argument (rather than capturing the current span) ([tokio-rs#1492]) Thanks to @CAD97 for contributing to this release! [v03]: https://github.com/tokio-rs/tracing/releases/tag/tracing-subscriber-0.3.0 [tokio-rs#635]: tokio-rs#635 [tokio-rs#695]: tokio-rs#695 [tokio-rs#1144]: tokio-rs#1144 [tokio-rs#1492]: tokio-rs#1492 [tokio-rs#1677]: tokio-rs#1677
No description provided.