perf(core): rework attribute serialisation

[43081j]

Does a couple of things:

• Switches to a regular for...in which is much faster than
  Object.entries
• Only retrieves the val if the key is valid
• Compares keys against "" (empty string) rather than accessing
  length
• Reuses typeof result rather than recomputing it each time

Type of change

Performance improvement

How Has This Been Tested?

Existing unit tests cover this code, and pass successfully.

Checklist:

• Followed the style guidelines of this project
• Unit tests have been added
• Documentation has been updated

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: 43081j / name: James Garbutt (25b00e4, 4403952, 6a0bf30, a11a360, ca82c05, dc7e48c)

[codecov]

Codecov Report

❌ Patch coverage is 92.30769% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 94.99%. Comparing base (91ff198) to head (25b00e4).
⚠️ Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
...ckages/opentelemetry-core/src/common/attributes.ts	92.30%	1 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@ ## main #5866 +/- ## ========================================== - Coverage 95.00% 94.99% -0.01%  ========================================== Files 313 313 Lines 8771 8775 +4 Branches 1883 1884 +1 ========================================== + Hits 8333 8336 +3  - Misses 438 439 +1 

Files with missing lines	Coverage Δ
...ckages/opentelemetry-core/src/common/attributes.ts	91.66% <92.30%> (-1.52%)	⬇️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[raphael-theriault-swi]

Could you provide some comparative figures given this is for perf ?

[43081j]

here's the results from a quick benchmark locally:

benchmark avg (min … max) p75 / p99 (min … top 1%) ------------------------------------------- ------------------------------- sanitizeAttributes 114.91 ns/iter 115.67 ns ▅█ (109.01 ns … 186.76 ns) 135.03 ns ██ (110.75 b … 611.93 b) 400.31 b ▁▁███▃▃▄▅▃▂▂▂▂▁▁▁▁▁▁▁ sanitizeAttributesMain 154.84 ns/iter 156.72 ns █ (150.32 ns … 251.54 ns) 170.29 ns █▃ (821.69 b … 2.59 kb) 1.00 kb ▆██▄▃▃▄▄▃▂▂▃▃▂▁▂▁▁▁▁▁ summary sanitizeAttributes 1.35x faster than sanitizeAttributesMain 

to be honest though, a large amount of the time is lost to warn calls 🤔

i wonder if you could optimise the log proxy to compute the logger up front?

basically this:

 const logger = getGlobal('diag'); if (!logger) return noop; const logFn = logger[funcName]; return logFn;

but it'd mean if you mutate globalThis later to add log functions, they wouldn't be picked up. so maybe not

[legendecas]

Thanks for the work! LGTM % a nit.

[david-luna]

@43081j thanks for your contribution. I think this change needs an entry into CAHNGELOG.md file

[43081j]

i've added an entry under "internal" if that works

[43081j]

we had a conflict from another merged PR, so I've caught the branch up again FYI

[otelbot]

Thank you for your contribution @43081j! 🎉 We would like to hear from you about your experience contributing to OpenTelemetry by taking a few minutes to fill out this survey.