fix: initialize agent user_id on agent arun

[JasonLovesDoggo]

Summary

Properly init Agent.user_id if passed in via Agent.arun

Type of change

• Bug fix
• New feature
• Breaking change
• Improvement
• Model update
• Other:

---

Checklist

• Code complies with style guidelines
• Ran format/validation scripts (./scripts/format.sh and ./scripts/validate.sh)
• Self-review completed
• Documentation updated (comments, docstrings)
• Examples and guides: Relevant cookbook examples have been included or updated (if applicable)
• Tested in clean environment
• Tests added/updated (if applicable)

---

Additional Notes

Add any important context (deployment instructions, screenshots, security considerations, etc.)
This was due to user_id not showing up in arize. We don't use teams.

[JasonLovesDoggo]

required as seen by https://github.com/Arize-ai/openinference/blob/main/python/instrumentation/openinference-instrumentation-agno/src/openinference/instrumentation/agno/_wrappers.py#L83

[JasonLovesDoggo]

Actually, just noticed the existing call was a self assignment. Probably a typo. It's initialized in Agent's init but not in arun. see https://github.com/agno-agi/agno/blob/main/libs/agno/agno/agent/agent.py#L445

[dirkbrnd]

@JasonLovesDoggo We don't actually want that to be "sticky" when passed on arun. It is supposed to only be for that run, otherwise it breaks multi-user flows.

[JasonLovesDoggo]

@JasonLovesDoggo We don't actually want that to be "sticky" when passed on arun. It is supposed to only be for that run, otherwise it breaks multi-user flows.

Yet session id persists?

I feel like I'm missing something here where we obviously want the telemetry to get the user id but it reads from an empty value.

The current code is just doing a self assignment which also just makes no sense

[dominic-mortimer]

Agreed!

[dirkbrnd]

@JasonLovesDoggo @dominic-mortimer
Agreed, this is not ideal for telemetry, but I believe the issue lies in the OpenInference package. Because the user_id is resolved in run and should then be correctly picked up downstream. But I see in the OpenInference package it uses the global value.

We don't want it to be sticky, because then it would break multi-user flows in production applications where an agent is used concurrently.

I'll look into updating the Arize package

[dirkbrnd]

Added a PR to openinference
Arize-ai/openinference#2138

I'll get it merged asap. Their team is usually pretty responsive

[JasonLovesDoggo]

Added a PR to openinference Arize-ai/openinference#2138

I'll get it merged asap. Their team is usually pretty responsive

Got it! Thanks Dirk. That's perfect for us