feat(anthropic): add Messages.create sync instrumentation #4034
+1,015
−67
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
github-actions
bot
assigned
codefromthecrypt,
karthikscale3,
lmolkova,
lzchen,
nirga and
vasantteja
- Removed unused trace import from conftest.py and updated tracer_provider fixture to return the provider instead of yielding it. - Enhanced test_instrument_with_no_providers to ensure proper instrument/uninstrument cycle while passing providers for a clean test environment.
- Improved readability by restructuring the assignment of request_model in the messages_create function. - Ensured consistent handling of model attributes from both the attributes dictionary and kwargs.
…s function - Updated the return type to include AttributeValue for better type safety. - Expanded the docstring to detail the OpenTelemetry semantic conventions for LLM requests and server attributes. - Included descriptions for GenAI attributes and server attributes to improve clarity and documentation.
vasantteja
force-pushed
the
feat/anthropic-messages-create-sync
branch
from
e011a91 to
643a282
Compare
- Introduced a new test file for integration tests that require a valid ANTHROPIC_API_KEY. - Implemented a pytest hook to skip integration tests if the API key is not set or is a placeholder. - Updated pyproject.toml to include pytest configuration for test paths and markers for integration tests. - Enhanced conftest.py to support skipping integration tests based on API key availability.
- Removed commented-out imports and improved assertion readability in test_integration.py. - Streamlined attribute checks for spans to enhance clarity and maintainability. - Ensured consistent validation of response attributes in integration tests.
- Added a pylint directive to disable the redefined-outer-name warning, improving code quality and maintainability.
- Updated the messages_create function to ensure request_model is explicitly converted to a string for better type consistency. - Cleaned up the docstring in get_llm_request_attributes to remove unnecessary whitespace and enhance readability. - Improved clarity in the documentation of GenAI and server attributes included in the returned dictionary.
- Added a comment to the AnthropicInstrumentor class to indicate the need for a logger_provider in the TelemetryHandler to capture content events.
Contributor
Author
|
@aabmass Can you please take another look? I updated the pr with the suggested changes. |
…nstrumentation package. The integration tests have been deleted, and the pytest configuration has been updated to reflect these changes.
…strumentation test configuration.
keith-decker
reviewed
Jan 9, 2026
Comment on lines
36
to
41
| def is_content_enabled() -> bool: | ||
| """Check if content capture is enabled via environment variable.""" | ||
| capture_content = environ.get( | ||
| OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT, "false" | ||
| ) | ||
| return capture_content.lower() == "true" |
Contributor
There was a problem hiding this comment.
This looks like something we should expose in genai utils if it's not already.
Looking through this PR, it doesn't look like it's actually used? I see it's passed to the wrapper function, but the _capture_content is not accessed. Can this just be removed and let GenAI utils take care of removing it from the spans/events?
…d related references. Updated messages_create function to eliminate unused parameter.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
9 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Implements OpenTelemetry instrumentation for Anthropic's
Messages.createAPI (sync, non-streaming). This adds automatic tracing for Anthropic SDK calls, capturing GenAI semantic convention attributes for observability.P.S: LLM help was used for writing the code.
What's Implemented
patch.py: Wrapper function forMessages.createthat creates spans with request/response attributesutils.py: Helper functions for attribute extraction, error handling, and content capture configuration__init__.py: Wiring for patching/unpatching viawraptSemantic Convention Attributes Captured
Request:
gen_ai.operation.name(chat)gen_ai.system(anthropic)gen_ai.request.modelgen_ai.request.max_tokensgen_ai.request.temperaturegen_ai.request.top_pgen_ai.request.top_kgen_ai.request.stop_sequencesserver.addressResponse:
gen_ai.response.idgen_ai.response.modelgen_ai.response.finish_reasonsgen_ai.usage.input_tokensgen_ai.usage.output_tokensError:
error.type(on exceptions)Fixes #(#3949)
Type of change
How Has This Been Tested?
Ran the full test suite with VCR cassettes for mocked API responses:
Test cases:
Results: 15 tests passed (8 new + 7 existing instrumentor tests)
Does This PR Require a Core Repo Change?
Checklist: