fix(openai-agents): capture response.instructions as system prompt on generation spans#3739
Open
ritiztambi wants to merge 1 commit intotraceloop:mainfrom
Open
Conversation
|
|
![ellipsis-dev[bot]](https://avatars.githubusercontent.com/in/64358?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Important
Looks good to me! 👍
Reviewed everything up to 9b3bdd2 in 12 seconds. Click for details.
- Reviewed
47lines of code in1files - Skipped
0files when reviewing. - Skipped posting
0draft comments. View those below. - Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.
Workflow ID: wflow_ZgSQMc8t1OACSjmc
You can customize 
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review infoConfiguration used: defaults Review profile: CHILL Plan: Pro Disabled knowledge base sources:
📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
📝 WalkthroughWalkthroughPrepend Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~12 minutes Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
🧹 Nitpick comments (1)
packages/opentelemetry-instrumentation-openai-agents/opentelemetry/instrumentation/openai_agents/_hooks.py (1)
626-637: Deduplicate the instruction-prepend block to avoid behavior drift.The same logic is duplicated in both branches. A small helper keeps both paths consistent and easier to maintain.
Refactor proposal
@@ def _extract_response_attributes(otel_span, response, trace_content: bool): @@ return model_settings + + +def _prepend_system_instruction(input_data, response, trace_content: bool): + if not ( + trace_content + and response + and hasattr(response, "instructions") + and response.instructions + ): + return input_data + system_msg = {"role": "system", "content": response.instructions} + if not input_data: + return [system_msg] + return [system_msg] + (input_data if isinstance(input_data, list) else list(input_data)) @@ - if ( - trace_content - and response - and hasattr(response, "instructions") - and response.instructions - ): - system_msg = {"role": "system", "content": response.instructions} - input_data = [system_msg] + (input_data if input_data else []) + input_data = _prepend_system_instruction( + input_data, response, trace_content + ) @@ - if ( - trace_content - and response - and hasattr(response, "instructions") - and response.instructions - ): - system_msg = {"role": "system", "content": response.instructions} - input_data = [system_msg] + (input_data if input_data else []) + input_data = _prepend_system_instruction( + input_data, response, trace_content + )Also applies to: 687-695
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@packages/opentelemetry-instrumentation-openai-agents/opentelemetry/instrumentation/openai_agents/_hooks.py` around lines 626 - 637, The duplicated block that prepends a system instruction to input_data (checking span_data.response.instructions and creating system_msg) should be extracted into a single helper function (e.g., _prepend_system_instructions(span_data, input_data) or _maybe_prepend_system_instructions) and invoked from both places in _hooks.py where the logic currently appears (the block around the response = getattr(span_data, "response", None) and the duplicate at 687-695); implement the helper to return the potentially modified input_data, handle None input_data correctly, and update the two call sites (instead of duplicating the condition and list construction) so behavior remains identical and easier to maintain.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In
`@packages/opentelemetry-instrumentation-openai-agents/opentelemetry/instrumentation/openai_agents/_hooks.py`:
- Around line 626-637: The duplicated block that prepends a system instruction
to input_data (checking span_data.response.instructions and creating system_msg)
should be extracted into a single helper function (e.g.,
_prepend_system_instructions(span_data, input_data) or
_maybe_prepend_system_instructions) and invoked from both places in _hooks.py
where the logic currently appears (the block around the response =
getattr(span_data, "response", None) and the duplicate at 687-695); implement
the helper to return the potentially modified input_data, handle None input_data
correctly, and update the two call sites (instead of duplicating the condition
and list construction) so behavior remains identical and easier to maintain.
ℹ️ Review info
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Disabled knowledge base sources:
- Linear integration is disabled
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (1)
packages/opentelemetry-instrumentation-openai-agents/opentelemetry/instrumentation/openai_agents/_hooks.py
… generation spans
ritiztambi
force-pushed
the
fix/openai-agents-capture-instructions
branch
from
9b3bdd2 to
4f2e321
Compare
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
2 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.
Fixes #3738
Problem
The
opentelemetry-instrumentation-openai-agentspackage does not captureresponse.instructionsfrom the OpenAI Responses API. When an agent hasinstructionsset (the system prompt), generation spans have nogen_ai.promptentry withrole: system— the system prompt is silently dropped.The
_extract_response_attributes()function readstemperature,max_output_tokens,top_p,model,frequency_penalty,output, andusage— but neverresponse.instructions.Fix
In
on_span_end, when handlingGenerationSpanData/ResponseSpanData, prependresponse.instructionsas a system message to the input data before calling_extract_prompt_attributes. This produces:gen_ai.prompt.0.role="system"gen_ai.prompt.0.content=<agent instructions>Followed by the conversation history at indices 1, 2, 3, etc.
Applied to both the primary code path (ResponseSpanData / GenerationSpanData) and the legacy fallback path.
Reference
This matches exactly how the vanilla
opentelemetry-instrumentation-openaipackage handles instructions inresponses_wrappers.py:Testing
Tested with an OpenAI Agents SDK agent that has
instructionsset. Before: no system prompt in generation spans. After: system prompt appears asgen_ai.prompt.0with rolesystem.Important
Capture
response.instructionsas a system prompt in generation spans inon_span_endin_hooks.py.on_span_endin_hooks.py, prependresponse.instructionsas a system message to input data forGenerationSpanDataandResponseSpanData.gen_ai.prompt.0.roleto"system"andgen_ai.prompt.0.contentto<agent instructions>.opentelemetry-instrumentation-openaiin handling instructions.instructionsset; system prompt now appears in generation spans.This description was created by
for 9b3bdd2. You can customize this summary. It will automatically update as commits are pushed.
Summary by CodeRabbit