v3.14.0rc1
📦 datadog-sdkView on GitHub →
✨ 3 features🐛 6 fixes🔧 1 symbols
Summary
This release introduces AI Guard support for LangChain and enhances LLM Observability by making `ml_app` optional and adding tool definition annotation capabilities. Several bug fixes address issues in asyncpg, exception replay, LLM observability, and CI Visibility.
✨ New Features
- ai_guard: add AI Guard evaluations support to LangChain instrumentation
- LLM Observability: `ml_app` is now optional, defaulting to `service`. Enabling LLM Observability will no longer throw if `ml_app` is not provided or propagated from an upstream service.
- LLM Observability: Introduces a `tool_definitions` parameter to the `LLMObs.annotate()` method for tool calling scenarios, allowing users to pass a list of tool definition dictionaries.
🐛 Bug Fixes
- asyncpg: Fixed an error related to custom connect options when using `asyncpg.connect`, ensuring postgres.connect spans are created.
- exception replay: Fixed an issue preventing snapshots from retrieving local variables from traceback frames in exceptions thrown by Celery tasks.
- LLM Observability: Resolved an issue where decorated functions returning responses with ambiguous truth values (e.g., pandas dataframes) raised an error during boolean coercion.
- Fixed an issue that could cause some products to fail startup in applications using `pkg_resources`.
- profiling: Upgraded echion to incorporate critical bug fixes and performance improvements.
- CI Visibility: Fixed an issue where coverage from pytest-xdist sessions was submitted without the proper session ID, impacting Test Impact Analysis.
🔧 Affected Symbols
LLMObs.annotate