Elevated error rates on OpenClaw through ChatGPT Subscription
Open 💬 5 comments Opened Mar 10, 2026 by rohannagpal
What issue are you seeing?
Seeing elevated rates of server_errors while using OpenClaw with my ChatGPT subscription. This is the message I get:
Codex error: {"type":"error","error":{"type":"server_error","code":"server_error","message":"An error occurred while processing your request. You can retry your request, or contact us through our help center at [help.openai.com](http://help.openai.com/) if the error persists. Please include the request ID e67376e4-d125-43aa-9698-8dccfe1c2171 in your message.","param":null},"sequence_number":2}
Here are three of the request IDs, all from the same session, I was told to use while contacting support:
- 6df14aa3-a6f4-4211-9c2a-480ff18d0df7
- a7c593a7-170f-4549-b59b-9255c16d0452
- e67376e4-d125-43aa-9698-8dccfe1c2171
What steps can reproduce the bug?
They just keep cropping up in long threads with my OpenClaw installation. Maybe its something to do with compaction?
What is the expected behavior?
_No response_
Additional information
I'm on a ChatGPT Pro account.
Ran into the most recent server_error at 1751 IST, 42 minutes after the degraded performance issue was marked resolved on status.openai.com.
This issue has 5 comments on GitHub. Read the full discussion on GitHub ↗