Codex VS Code WSL app-server stuck at 70–90% CPU with outbound queue is full

Open 💬 3 comments Opened May 1, 2026 by mkikunda

Hello,

I am using Codex VS Code extension with ChatGPT Pro on Windows + WSL Ubuntu.

Codex is installed only in WSL: Ubuntu, not locally anymore.

Extension version:
openai.chatgpt 26.422.71525

Problem:
As soon as I open the Codex panel, the local Codex app-server starts consuming 70–90% CPU and the panel becomes unusable.

Process shown in WSL:
~/.vscode-server/extensions/openai.chatgpt-26.422.71525-linux-x64/bin/linux-x86_64/codex app-server --analytics-default-enabled

Logs:
codex_app_server::transport: dropping overload response for connection ConnectionId(0):
outbound queue is full

Sometimes the process remains as:
[codex] <defunct>

I already tried:

  • uninstalling and reinstalling the extension
  • removing the local installation and keeping only WSL
  • killing codex app-server
  • wsl --shutdown
  • restarting VS Code
  • reconnecting/reloading the window

The issue persists after reinstall.

Please advise or provide a workaround/rollback version.

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗