One or more namespaces failed to connect... but the list is empty

[ffteja]

I have this traceback.
This happens very rarely but this is the second time I observe it.

... pip list
[2025-11-14T18:10:34Z]: Package           Version
[2025-11-14T18:10:34Z]: ----------------- -------
[2025-11-14T18:10:34Z]: aiohappyeyeballs  2.6.1
[2025-11-14T18:10:34Z]: aiohttp           3.13.2
[2025-11-14T18:10:34Z]: aiosignal         1.4.0
[2025-11-14T18:10:34Z]: attrs             25.4.0
[2025-11-14T18:10:34Z]: bidict            0.23.1
[2025-11-14T18:10:34Z]: frozenlist        1.8.0
[2025-11-14T18:10:34Z]: h11               0.16.0
[2025-11-14T18:10:34Z]: idna              3.11
[2025-11-14T18:10:34Z]: multidict         6.7.0
[2025-11-14T18:10:34Z]: pip               25.3
[2025-11-14T18:10:34Z]: propcache         0.4.1
[2025-11-14T18:10:34Z]: python-engineio   4.12.3
[2025-11-14T18:10:34Z]: python-socketio   5.14.3
[2025-11-14T18:10:34Z]: setuptools        65.5.1
[2025-11-14T18:10:34Z]: simple-websocket  1.1.0
[2025-11-14T18:10:34Z]: typing_extensions 4.15.0
[2025-11-14T18:10:34Z]: wsproto           1.3.1
[2025-11-14T18:10:34Z]: yarl              1.22.0
... execution
[2025-11-14T18:11:36Z]: Traceback (most recent call last):
[2025-11-14T18:11:36Z]:   File "/enc/agent/main.py", line 1240, in run
[2025-11-14T18:11:36Z]:     await sio.connect(host, auth=auth, socketio_path=url_path + "/api/v1/socket.io", retry=True)
[2025-11-14T18:11:36Z]:   File "/enc/agent/pythonnca/lib64/python3.11/site-packages/socketio/async_client.py", line 177, in connect
[2025-11-14T18:11:36Z]:     raise exceptions.ConnectionError(
[2025-11-14T18:11:36Z]: socketio.exceptions.ConnectionError: One or more namespaces failed to connect: 
[2025-11-14T18:11:36Z]: 
[2025-11-14T18:11:37Z]: 2 logs submitted to AppInsights

I don't understand the error message.
This is the source code:

python-socketio/src/socketio/async_client.py

Line 175 in db3f1c2

if set(self.namespaces) != set(self.connection_namespaces):

Questions:
1/ If "self.failed_namespaces" is empty, what could be the root cause ? My suspicion is the TimeoutError before any namespace could be put into self.failed_namespaces collection. Then the error message would be a bit misleading.
2/ If the initial connection can fail should I retry myself (the retry is only valid after initial connection?)

[miguelgrinberg]

The message is slightly misleading. It should be something like "One or more namespaces timed out or failed to connect.". The failed namespaces are already collected and can be shown directly as it is done now. The timed out namespaces have to be calculated as those that are not connected or failed.

1/ If "self.failed_namespaces" is empty, what could be the root cause ? My suspicion is the TimeoutError before any namespace could be put into self.failed_namespaces collection. Then the error message would be a bit misleading.

Your issue is that one or more namespaces timed out. You can enable logging to have more info about what happened.

2/ If the initial connection can fail should I retry myself (the retry is only valid after initial connection?)

It probably makes more sense to figure out why your namespaces are timing out, because it would be unusual that the Engine.IO connection works just fine but then some of the namespaces time out.