Skip to content

feat: Updated Agno attributes #2262

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 11 commits into
base: main
Choose a base branch
from
Open

feat: Updated Agno attributes #2262

wants to merge 11 commits into from
+298 −60

Conversation

willemcdejongh
Copy link

@willemcdejongh willemcdejongh commented Oct 2, 2025 •
edited by cursor bot
Loading

Note

Upgrade to Agno v2 and update wrappers to extract cleaner inputs/outputs, aggregate streamed results, and record LLM token/cache metrics with refined message attributes.

  • Agno instrumentation:
    • Dependency: Bump agno to >=2.1.2 (tests pin ==2.1.2).
    • Agent spans:
      • Set input.value from user message content via RunOutput/RunMessages (_get_user_message_content).
      • Set output.value from RunOutput.content (_extract_run_response_output).
      • Improve streaming handling: read from session.runs or agent.get_last_run_output; ensure output.mime_type=application/json and OK status.
    • LLM spans:
      • Input: restrict to cleaned messages only; limit roles to system|user|assistant in llm.input_messages.
      • Output: parse responses to populate llm.output_messages; aggregate streamed chunks into a single message (_parse_model_output_stream).
      • Metrics: record llm.token_count.prompt, llm.token_count.completion, and cache read/write tokens on sync/async (new span attributes added).
    • Helpers/attrs: Add _parse_model_output, _parse_model_output_stream, _extract_run_response_output; define token/cache span attribute constants.

Written by Cursor Bugbot for commit fc65734. This will update automatically on new commits. Configure here.

@willemcdejongh willemcdejongh requested a review from a team as a code owner October 2, 2025 11:24
@dosubot dosubot bot added the size:M This PR changes 30-99 lines, ignoring generated files. label Oct 2, 2025
@github-actions GitHub Actions
Copy link
Contributor

github-actions bot commented Oct 2, 2025 •
edited
Loading

CLA Assistant Lite bot All contributors have signed the CLA ✍️ ✅

cursor[bot]

This comment was marked as outdated.

Sign in to view

@dosubot dosubot bot added size:L This PR changes 100-499 lines, ignoring generated files. and removed size:M This PR changes 30-99 lines, ignoring generated files. labels Oct 6, 2025
cursor[bot]

This comment was marked as outdated.

Sign in to view

cursor[bot]

This comment was marked as outdated.

Sign in to view

@willemcdejongh
Copy link
Author

I have read the CLA Document and I hereby sign the CLA

github-actions bot added a commit that referenced this pull request Oct 8, 2025
@github-actions
@willemcdejongh has signed the CLA in #2262
7bebd6b
@willemcdejongh willemcdejongh changed the title DRAFT: feat: Updated Agno attributes feat: Updated Agno attributes Oct 8, 2025
cursor[bot]

This comment was marked as outdated.

Sign in to view

cursor[bot]

This comment was marked as outdated.

Sign in to view

cursor[bot]

This comment was marked as outdated.

Sign in to view

cursor[bot]

This comment was marked as outdated.

Sign in to view

cursor[bot]

This comment was marked as outdated.

Sign in to view

dirkbrnd
dirkbrnd reviewed Oct 8, 2025
View reviewed changes
...ation/openinference-instrumentation-agno/src/openinference/instrumentation/agno/_wrappers.py
run_response = agent.get_last_run_output(session_id=session_id)
if "session" in arguments and arguments.get("session") and len(arguments.get("session").runs) > 0:
for run in arguments.get("session").runs:
if run.content:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But if the session has many runs, this won't work? You'll have to get the last run in the session.

dirkbrnd
dirkbrnd reviewed Oct 8, 2025
View reviewed changes
...ation/openinference-instrumentation-agno/src/openinference/instrumentation/agno/_wrappers.py
if run_response is not None:
span.set_attribute(OUTPUT_VALUE, run_response.to_json())
span.set_attribute(OUTPUT_MIME_TYPE, JSON)
else:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this Else necessary?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Its a last resort fallback. Users that have caching or a DB enabled, might still get a response here.

dirkbrnd
dirkbrnd reviewed Oct 8, 2025
View reviewed changes
...ation/openinference-instrumentation-agno/src/openinference/instrumentation/agno/_wrappers.py
for i, message in enumerate(messages):
role, content = message.role, message.get_content_string()
if content:
if role in ["system", "user", "assistant"]:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What about tool results messages?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is on input. We map the tool results in the output messages

dirkbrnd
dirkbrnd reviewed Oct 8, 2025
View reviewed changes
...ation/openinference-instrumentation-agno/src/openinference/instrumentation/agno/_wrappers.py Outdated
else:
return str(output)

if hasattr(output, 'role') and hasattr(output, 'content') and hasattr(output, 'tool_calls'):
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Some output would have only content, and some only tool_calls. Should it not be or?

dirkbrnd
dirkbrnd reviewed Oct 8, 2025
View reviewed changes
...ation/openinference-instrumentation-agno/src/openinference/instrumentation/agno/_wrappers.py Outdated
"role": getattr(output, 'role', None),
"content": getattr(output, 'content', None),
"tool_calls": getattr(output, 'tool_calls', []),
"event": getattr(output, 'event', None),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't think we should include this.

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

removed

dirkbrnd
dirkbrnd reviewed Oct 8, 2025
View reviewed changes
...ation/openinference-instrumentation-agno/src/openinference/instrumentation/agno/_wrappers.py

messages.append(result_dict)

return {"messages": messages}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should it return like this? Should it not just be a single string as the output?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We do a json dumps with the response dict right?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@graphite-app graphite-app[bot] graphite-app[bot] left review comments

@cursor cursor[bot] cursor[bot] left review comments

+1 more reviewer

@dirkbrnd dirkbrnd dirkbrnd left review comments

Reviewers whose approvals may not affect merge requirements

At least 1 approving review is required to merge this pull request.

Assignees
No one assigned
Labels
size:L This PR changes 100-499 lines, ignoring generated files.
Projects
Instrumentation
Status: No status
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@willemcdejongh @dirkbrnd