Arize AI Community Icon

Home
Events

Arize AX Releases
Arize AX Support
Arize News
Arize Observe
Discussions
Introductions
Phoenix Support

Powered by Tightknit

Phoenix Support

Troubleshooting Google ADK Error with Phoenix Cloud Integration | Arize AI Community

Phoenix Support

Troubleshooting Google ADK Error with Phoenix Cloud Integration

·Jul 30, 2025 09:21 PM

Hello, I am using phoenix cloud and starting using it, so far the information looks good inside Phoenix, however since I am using Google ADK with Phoenix, when I am running it I keep getting the same error:

Traceback (most recent call last):
  File "C:\Users\Tomy\anaconda3\envs\google_adk\lib\site-packages\opentelemetry\context\__init__.py", line 155, in detach
    _RUNTIME_CONTEXT.detach(token)
  File "C:\Users\Tomy\anaconda3\envs\google_adk\lib\site-packages\opentelemetry\context\contextvars_context.py", line 53, in detach
    self._current_context.reset(token)
ValueError: <Token var=<ContextVar name='current_context' default={} at 0x000001C65274E7A0> at 0x000001C6023F0F80> was created in a different Context

I am not even sure if this is something to review here or if I need to review it with google adk by itself, just wondering if anyone has seen a similar problem and have fixed it before?

11 comments

· Sorted by Oldest

RunLLM
APP
·
This error is a known issue when using Phoenix (and OpenInference) with OpenTelemetry 1.34.0+ in async code, especially with Google ADK. It occurs because context tokens are being detached in a different async context than where they were created, often due to how async generators and context managers interact in Python and the ADK SDK. The error is usually harmless but can be avoided by downgrading OpenTelemetry to 1.33.1 or by manually configuring OpenTelemetry instead of using Phoenix's register helper. See detailed discussion and workarounds in Phoenix issue #8030 and OpenInference issue #1769.Would you like more detail on how to apply these workarounds or why this happens?Sources: - Phoenix issue #8030 - OpenInference issue #1769
💡 Hint: Mention RunLLM in the thread for followups.
Xander S.
·
Hey Tomas B., are you able to share a snippet to reproduce the issue? This error sort of looks like it might have something to do with context variables being instantiated in a non-local context (e.g., if you use multi-threading in your application code).
Roger Y.
·
may be related to this, which we don’t have a good fix for
Roger Y.
·
but if you can give us the code to reproduce it, we’ll investigate
Tomas B.
·
Xander S. yes I am using threading, I have pretty much a main.py that creates different threads for chatting with the ai agent using, slack, discord, api and cli, however even if I remove the threads I am still getting the error
Roger Y.
·
do you get similar error when not applying the instrumentor?
Tomas B.
·
but, yes it's similar to that one too Roger Y., same error that I am seeing, it's annoying because even the data in Phoenix looks good,
Tomas B.
·
Roger Y. hmm, not sure what you mean with instrumentor, I am using directly:
from phoenix.otel import register tracer_provider = register( project_name="my-project", auto_instrument=True )
auto_instrument = true or false == same error in my case from what I remember
Roger Y.
·
ok got it
Tomas B.
·
let me see if I can come up with code tomorrow, so you can check it out too, I am open to work arounds at this point haha.
👍1
Roger Y.
·
sounds good. thanks