Repository navigation
RuntimeError: can't create new thread at interpreter shutdown #4752
Description
Describe your environment
OS: Ubuntu
Python version: 3.12.0, 3.12.1, 3.12.2
SDK version: Since 1.36.0
API version: Since 1.36.0
What happened?
Exception ignored in atexit callback: <function shutdown at 0x72be9c2a93a0>
Traceback (most recent call last):
File "/home/codespace/.python/current/lib/python3.12/logging/__init__.py", line 2271, in shutdown
h.flush()
File "/workspaces/opentelemetry-python/opentelemetry-sdk/src/opentelemetry/sdk/_logs/_internal/__init__.py", line 654, in flush
thread.start()
File "/home/codespace/.python/current/lib/python3.12/threading.py", line 992, in start
_start_new_thread(self._bootstrap, ())
RuntimeError: can't create new thread at interpreter shutdown
Steps to Reproduce
- Run the example here with python 3.12.0
Expected Result
No RuntimeError
Actual Result
RuntimeError
Additional context
Got this after this PR
Probably due to: python/cpython#113964
Would you like to implement a fix?
None
Tip
React with 👍 to help prioritize this issue. Please use comments to provide useful context, avoiding +1 or me too, to help us triage it. Learn more here.
Reacted by jetunc and Jarrad
Activity
emdneto commentedon Sep 16, 2025
From python >=3.12.3 there's no issue. it seems the issue happens with python 3.12.0, 3.12.1 and 3.12.2
aabmass commentedon Oct 9, 2025
Is this WAI @emdneto ? Last time I looked at this the CPython bugs insisted this is the expected behavior
emdneto commentedon Dec 24, 2025
Hey @aabmass, sorry, I missed this.
It seems so, but we still see the RuntimeError on our end. AFAICS, people are catching the error and suppressing it as a workaround.
fix(sdk): handle RuntimeError during LoggingHandler.flush at shutdown
fix(sdk): handle RuntimeError during LoggingHandler.flush at shutdown