code and a type. The type is a URL that resolves to a page on this site, so an error message can point at its own explanation.
The base is https://docs.cortex.foundation/problems. Each code has a page at /problems/{code}, and the path is the snake_case code — rate_limited, not rate-limited.
Start with the status
Retryable: yes means the same request can succeed later. Retryable: no means retrying changes nothing — the request, the plan, or the state has to change first.
All codes
Two 429s that mean different things
rate_limited is transport throttling — you are sending requests too fast, and waiting fixes it. quota_exceeded is a plan window running out — waiting for the reset or changing plan fixes it. Read the code, not the status.
Related
- Errors - the shape of a problem document and which fields to render.
- Platform API - why there is no public inference API to raise these codes against yet.
- Troubleshooting - what a message in the product means.