fix(llmobs): safely handle non-json serializable arguments [backport #10694 to 2.12]

[Yun-Kim]

Backports #10694 to 2.12.

Safely handle non-JSON serializable tag arguments in LLMObs.annotate() and the OpenAI/LangChain/Bedrock/Anthropic integrations.

• LLMObs.annotate(): we previously just discarded the entire argument to annotate if the argument was non-JSON serializable. Now, we safely convert non-JSON serializable fields/objects to a default placeholder text, meaning users can still send some data even if some of it may be invalid.
• Same idea with each integration, we ensure we safely handle non-JSON serializable args and default to placeholder texts if necessary.
• We've moved all json.dumps() call into a private helper safe_json() which does the above for us.

Note: This PR removes some tests in test_llmobs_service.py regarding truly unserializable objects as this is highly unlikely, someone would have to go out of their way to make a truly unserializable object (i.e. override __repr__() with a non-json serializable value). We still catch these so any resulting crashes should not be from our code.

Checklist

• PR author has checked that all the criteria below are met
• The PR description includes an overview of the change
• The PR description articulates the motivation for the change
• The change includes tests OR the PR description describes a testing strategy
• The PR description notes risks associated with the change, if any
• Newly-added code is easy to change
• The change follows the library release note guidelines
• The change includes or references documentation updates if necessary
• Backport labels are set (if applicable)

Reviewer Checklist

• Reviewer has checked that all the criteria below are met
• Title is accurate
• All changes are related to the pull request's stated goal
• Avoids breaking API changes
• Testing strategy adequately addresses listed risks
• Newly-added code is easy to change
• Release note makes sense to a user of the library
• If necessary, author has acknowledged and discussed the performance implications of this PR as reported in the benchmarks PR comment
• Backport labels are set in a manner that is consistent with the release branch maintenance policy

[pr-commenter]

Benchmarks

Benchmark execution time: 2024-09-20 22:21:20

Comparing candidate commit d99a5ae in PR branch yunkim/backport-10694-212 with baseline commit 845719b in branch 2.12.

Found 0 performance improvements and 0 performance regressions! Performance is the same for 353 metrics, 47 unstable metrics.

[datadog-dd-trace-py-rkomorn]

Datadog Report

Branch report: yunkim/backport-10694-212
Commit report: d99a5ae
Test service: dd-trace-py

✅ 0 Failed, 10445 Passed, 51738 Skipped, 3h 23m 43.38s Total duration (4h 38m 57.77s time saved)