Repository navigation
Automatically retry when an OpenAI model is at capacity #8570
Description
Area
Provider adapters
What are you trying to accomplish?
I want Codex Desktop and Codex CLI to automatically retry requests when the OpenAI model is at capacity, instead of having to click Retry manually.
What prevents this today?
Every time the model is at capacity, Codex stops processing my request, and I have to retry it manually.
What should OpenCodex do?
When the OpenAI model is at capacity, Codex should automatically retry the request with a short delay and jitter until the model becomes available, without interrupting the task.
Example usage or interface
Alternatives or workarounds
No response
Additional context
No response
Checks
- I searched existing issues and documentation.
- This request describes a concrete OpenCodex workflow rather than merely naming a desired technology.
- I removed secrets and personal data.
Activity
github-actions commentedon Nov 11, 2026
Issue closed: insufficient detail
Thanks for taking the time to submit this issue. It was closed automatically because the structured report is missing information needed to evaluate or reproduce it.
Please edit the issue to add:
See the Contributing guide for details.
Once the report passes the automated checks, it will be reopened automatically unless a maintainer has changed its state.