Describe your environment

We have a Python 3.11 app using openllmetry 0.38.8

What happened?

We encountered this bug in openllmetry - traceloop/openllmetry#2699

 File "/code/python/service/ai_agents/.venv/lib/python3.11/site-packages/opentelemetry/sdk/trace/__init__.py", line 713, in _from_env_if_absent
    raise ValueError(err_msg.format(env_var, value))
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: cannot switch from manual field specification to automatic field numbering

We still don't know what's causing the original problem, but the error itself is dubious. We should seen a different error that tells us which env var that has the malformed value, but instead run into a ValueError on the error message formatting.

When I checked, it felt like the problem might be on this line -

1. Unless I've missed something, It feels like the formatting may be the culprit for the ValueError.
2. It's not clear how _from_env_if_absent is even failing in openllmetry, but I suppose fixing the above bug will help us reach that part. We have a gunicorn FastAPI app backing an LLM service in our setup.

Steps to Reproduce

Even a basic python script like the following reproduces the error:

env_var = "test"
value = -1
err_msg = "{0} must be a non-negative integer but got {}"
raise ValueError(err_msg.format(env_var, value))

Expected Result

Should tell us which env var was supplied an incorrect value

Actual Result

ValueError: cannot switch from manual field specification to automatic field numbering

Additional context

No response

Would you like to implement a fix?

None