Reconnecting... Codex CLI 0.63.0
Resolved 💬 8 comments Opened Nov 29, 2025 by Asentrixx Closed Nov 30, 2025
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
What version of Codex is running?
0.63.0
What subscription do you have?
API
Which model were you using?
gpt-5.1-codex
What platform is your computer?
Ubuntu 22.04
What issue are you seeing?
Constant timeouts and "reconnecting ... 1/5" until failure.
This has been happening for a while now.
I don't use a VPN or anything that could interfere with the network
If I interrupt the reconnecting and say something to the AI, it will respond without issues.
The context window is at 95%, I've tried clearing the chat as well and it doesn't seem to matter...
<img width="485" height="275" alt="Image" src="https://github.com/user-attachments/assets/ee2fe0ee-e0bc-4c4d-a0a0-5b0161d7032d" />
What steps can reproduce the bug?
Not sure since I don't know what's causing it?
What is the expected behavior?
The request should be fulfilled without disconnection issues
Additional information
_No response_
8 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Considering cancelling ChatGPT subscription 1/5
same...
Check ~/.codex/sessions, inside youll see the month folder, click the current month, then youll see day folders, check the properties of the one the day the issue started. It is likely multiple gb, if it is open that folder and find the jsonl that is multiple gb and delete it. If its not check the others for an abnormally large folder, find the abnormally large jsonl and delete.
the same here. I tried reinstall, logout, login, dancing, walking, running, sleeping, eating and so on
Thanks for reporting the problem — and for your patience as the team investigated the underlying cause.
We think issue is understood and mitigated. It appears to have been a bad routing rule in one server pod. That's why it affected only a small fraction of codex users and only sporadically.
The "Reconnecting..." message in the client is used for all general errors that the client thinks are transient. That's why it retries up to five times before giving up on such errors. We're considering ways to improve or augment the "Reconnecting..." error message so it's easier to diagnose transient issues in the future.
I'm going to close this issue as resolved. If you see "Reconnecting..." errors in the future, please open a new issue rather than commenting on this one because it's almost definitely a different underlying problem.
Still happening regularly, issue hasn't been resolved at all...
As for making a "new issue" it's absolutely still related to the same thing.
Nothing has changed, I stopped over a week ago, came back and the same problem persists.
<img width="704" height="246" alt="Image" src="https://github.com/user-attachments/assets/be6871cd-6790-4efc-83e8-fcc68f0d21aa" />
@Asentrixx, as I requested above, please open a new issue with details if you are seeing this.
I'm going to lock this issue to prevent additional posts.