tracing-opentelemetry examples doesn't work

[Sherlock-Holo]

Bug Report

Version

│   │   ├── tracing v0.1.22
│   │   │   ├── tracing-attributes v0.1.13 (proc-macro)
│   │   │   └── tracing-core v0.1.17
│   │   └── tracing-futures v0.2.4
│   │       └── tracing v0.1.22 (*)
│   ├── tracing v0.1.22 (*)
│   └── tracing v0.1.22 (*)
├── tracing v0.1.22 (*)
├── tracing-opentelemetry v0.11.0
│   ├── tracing v0.1.22 (*)
│   ├── tracing-core v0.1.17 (*)
│   ├── tracing-log v0.1.2
│   │   └── tracing-core v0.1.17 (*)
│   └── tracing-subscriber v0.2.16
│       ├── tracing v0.1.22 (*)
│       ├── tracing-core v0.1.17 (*)
│       ├── tracing-log v0.1.2 (*)
│       └── tracing-serde v0.1.2
│           └── tracing-core v0.1.17 (*)
├── tracing-subscriber v0.2.16 (*)
    ├── tracing v0.1.22 (*)
    └── tracing-futures v0.2.4 (*)

Platform

Archlinux

Crates

tracing-opentelemetry

Description

the examples in https://github.com/tokio-rs/tracing/tree/master/tracing-opentelemetry doesn't work, jaeger doesn't see the trace data

https://github.com/tokio-rs/tracing/blob/master/examples/examples/opentelemetry.rs this examples should be the right way

[jtescher]

@Sherlock-Holo the docs could probably be improved, there are basically two steps when configuring opentelemetry. 1) choose your exporter, the example you linked to shows a stdout exporter, other valid exporters include a jaeger exporter and there are many other options for exporting span data. and 2) configuring the opentelemetry sdk (specifying span sampling logic, trace and span id generation, etc).

Jaeger is a very common destination for sending span data, so perhaps it makes sense to have that be the first example that people see when the view the readme, the current example has the benefit of being simple and not showing the additional complexity, but as you point out it has the downside of not actually sending the span data anywhere which is not likely to be what people want to do.