feat: @google/genai Gemini AI LLM instrumentation

[amychisholm03]

This PR closes #2890. This instruments the @google/genai library which provides an SDK for Google's Gemini AI.

For reviewers, refer to:

• Python Agent Gemini Instrumentation CDD
• LLM Agent Spec
• Run the example app here
• @google/genai repo

[codecov]

Codecov Report

Attention: Patch coverage is 98.24176% with 8 lines in your changes missing coverage. Please review.

Project coverage is 80.27%. Comparing base (686c747) to head (eaa4c4c).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
lib/instrumentation/@google/genai.js	98.96%	3 Missing ⚠️
lib/llm-events/google-genai/embedding.js	88.00%	3 Missing ⚠️
lib/llm-events/error-message.js	88.23%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3119      +/-   ##
==========================================
- Coverage   80.27%   80.27%   -0.01%     
==========================================
  Files         322      328       +6     
  Lines       49103    49558     +455     
==========================================
+ Hits        39419    39781     +362     
- Misses       9684     9777      +93     

Flag	Coverage Δ
integration-tests-esm-20.x	49.64% <29.16%> (?)
integration-tests-esm-22.x	49.70% <29.16%> (-0.01%)	⬇️
versioned-tests-18.x	80.17% <98.24%> (?)
versioned-tests-20.x	80.18% <98.24%> (?)
versioned-tests-22.x	80.19% <98.24%> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[jsumners-nr]

It'll take some time to go through the actual code. But I wanted to get this style feedback over sooner rather than later.

[mrickard]

...though if we define parsedError from cause.message, we'll need to change this reference to parsedError.code

[amychisholm03]

I think it's better to set parsedError to undefined that way we don't attempt to set http.statusCode or error.code if that value doesn't actually exist.

[mrickard]

And it's true that we're capturing cause.message in error.message.

I'd recommend we add one test of a malformed cause.message string in test/versioned/google-genai/mock-responses.js, just to make sure that this error handling does what we think it does.

[amychisholm03]

Yup I added that test in test/unit/llm-events/google-genai/error.test.js per James' suggestion