Support OTel gen_ai semconv v1.40#109
Merged
krisztianfekete merged 3 commits intomainfrom Apr 7, 2026
Merged
Conversation
There was a problem hiding this comment.
Pull request overview
Adds end-to-end support for richer OTel GenAI semantic conventions (up to v1.40.0) by extracting additional gen_ai.* attributes in the backend and surfacing them in the UI (provider, response model, cache token usage, etc.), improving trace/session observability.
Changes:
- Backend: centralizes new GenAI attribute keys and expands extraction of provider/response metadata, cache token usage, finish reasons, error types, and request params.
- UI: extends trace/session/inspector types and rendering to display provider/response model and cache token metrics.
- Examples/deps: updates OpenAI-related dependencies and sets semconv stability opt-in for the OpenAI Agents example.
Reviewed changes
Copilot reviewed 15 out of 16 changed files in this pull request and generated 4 comments.
Show a summary per file
| File | Description |
|---|---|
| uv.lock | Bumps openai dependency version. |
| ui/src/lib/types.ts | Adds provider/responseModel/agentId fields and cache token metrics types. |
| ui/src/context/TraceProvider.tsx | Propagates new metadata fields into table rows during streaming evaluation. |
| ui/src/components/streaming/SessionMetadata.tsx | Refactors metadata rendering and adds cache token display (currently has an unused local). |
| ui/src/components/streaming/SessionCard.tsx | Displays provider badge and cache token summary; passes invocations into metadata panel. |
| ui/src/components/inspector/PerformanceSection.tsx | Displays provider/response model and cache token metrics in the inspector performance table. |
| ui/src/components/inspector/InspectorView.tsx | Passes trace info (provider/model/responseModel) into the comparison/performance panel. |
| ui/src/components/inspector/ComparisonPanel.tsx | Threads traceInfo through to PerformanceSection. |
| ui/src/api/client.ts | Maps streamed traceMetadata fields into partial TraceResult. |
| src/agentevals/trace_metrics.py | Extracts provider/response model and cache token totals into trace metadata/metrics. |
| src/agentevals/trace_attrs.py | Adds new GenAI semconv attribute constants (v1.40.0 coverage). |
| src/agentevals/streaming/ws_server.py | Extracts richer per-invocation model info from spans (provider/response models/finish reasons/cache tokens/etc.). |
| src/agentevals/extraction.py | Adds extract_extended_model_info_from_attrs + tool type/description extraction. |
| src/agentevals/api/models.py | Extends TraceConversionMetadata with agentId/provider/responseModel. |
| examples/zero-code-examples/openai-agents/run.py | Enables OTEL_SEMCONV_STABILITY_OPT_IN=gen_ai_latest_experimental. |
| examples/zero-code-examples/openai-agents/requirements.txt | Updates example requirements for OpenAI/OpenAI Agents. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
peterj
reviewed
Apr 7, 2026
src/agentevals/extraction.py
Outdated
| elif finish_reasons_raw: | ||
| finish_reasons = [finish_reasons_raw] | ||
|
|
||
| temperature = attrs.get(OTEL_GENAI_REQUEST_TEMPERATURE) |
Contributor
There was a problem hiding this comment.
a lot of repeated functional;ity, consider creating a helper that tries to cast to a provided type and returns the value (or default value)
peterj
reviewed
Apr 7, 2026
src/agentevals/extraction.py
Outdated
| for provider when gen_ai.provider.name is absent (backward compat with | ||
| pre-v1.37.0 instrumentors). | ||
| """ | ||
| provider = attrs.get(OTEL_GENAI_PROVIDER_NAME) |
Contributor
There was a problem hiding this comment.
provider = attrs.get(OTEL_GENAI_PROVIDERNAME) or attrs.get(OTEL_GENAI_SYSTEM)
peterj
reviewed
Apr 7, 2026
src/agentevals/extraction.py
Outdated
| if not provider: | ||
| provider = attrs.get(OTEL_GENAI_SYSTEM) | ||
|
|
||
| finish_reasons_raw = attrs.get(OTEL_GENAI_RESPONSE_FINISH_REASONS) |
Contributor
There was a problem hiding this comment.
extract this to a separate function (e.g. _parse_finish_reasons)
peterj
reviewed
Apr 7, 2026
| except (TypeError, ValueError): | ||
| cache_read = 0 | ||
|
|
||
| return { |
Contributor
There was a problem hiding this comment.
consider having a TypedDict for this
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This PR adds support for OTel GenAI semantic conventions v1.40.0 to the trace consumer pipeline. Previously we only extracted a handful of
gen_ai.*attributes from incoming traces. Now we pull out provider identity, response model metadata, finish reasons, cache token usage, tool type/description, error classification, and request parameters like temperature and max tokens when instrumentors include them.The practical effect is richer trace analysis out of the box. Sessions now show which provider served each request, the actual model that responded (which can differ from what was requested), and prompt caching stats for cost visibility.
Fixes #106