confighttp: add support for binding unix sockets

[axw]

Description

Embed confignet.AddrConfig in confighttp.ServerConfig, replacing the Endpoint field. This introduces the "transport" config attribute, which can be set to "unix" for binding a Unix domain socket.

Link to tracking issue

Fixes #8752
Fixes #14187

Testing

$ ./bin/otelcorecol_linux_amd64 \
    --config examples/local/otel-config.yaml \
    --set receivers::otlp::protocols::http::transport=unix \
    --set receivers::otlp::protocols::http::endpoint=\"@otlphttp\"

$ curl --abstract-unix-socket otlphttp http://testing.invalid:1234/v1/traces
405 method not allowed, supported: [POST]

Documentation

Updated README.md

[codspeed-hq]

CodSpeed Performance Report

Merging this PR will not alter performance

_{Comparing axw:confighttp-addrconfig (b7cc95a) with main (1e5f5cf)}

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

Summary

✅ 64 untouched benchmarks
⏩ 20 skipped benchmarks¹

Footnotes

1. 20 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.77%. Comparing base (fbd5d3f) to head (b7cc95a).
⚠️ Report is 7 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #14248      +/-   ##
==========================================
- Coverage   91.79%   91.77%   -0.02%     
==========================================
  Files         676      677       +1     
  Lines       42419    42433      +14     
==========================================
+ Hits        38938    38943       +5     
- Misses       2422     2428       +6     
- Partials     1059     1062       +3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[axw]

This is a breaking change for the API, but not for user configuration. Naturally there are a bunch of compilation errors in contrib, which I can fix if/when this is accepted.

[axw]

@dmitryax and I discussed this in a recent Collector SIG meeting, and Dmitry wanted me to look into phasing out Endpoint so it wouldn't be immediately breaking. I can't see a way to do this without breaking (API) changes.

Some components are directly constructing confighttp.ServerConfig, rather than using confighttp.NewDefaultServerConfig, so there's no way to guarantee that the NetAddr field is initialised. Then the Validate method on the NetAddr field fails because "" is not a valid transport.

I thought of adding an Unmarshal method to ServerConfig, but that doesn't work because ServerConfig is typically if not always squashed: see #13273 (comment)

I also thought of making the NetAddr field optional with configoptional, so it wouldn't cause a validation error when uninitialised, but that would be a bandaid fix; it wouldn't make sense for the field to be optional once the Endpoint field is removed.

So I think we should keep it simple, and make a breaking change and just fix up the contrib components immediately after.

[dmitryax]

This change looks good to me as is. Especially given that it brings consistency with the configgrpc. I'll wait for another approval from @open-telemetry/collector-approvers before merging

[atoulme]

Changes look ok to me. I also was in the SIG meeting of 12/9 when we discussed this. The intent behind this PR makes sense.

[mx-psi]

cc @open-telemetry/collector-approvers I will merge this by Wednesday unless there are comments by then