429 errors - quota vs rate limit

I noticed I get a 429 error for both calling too fast and exceeding the quota.

The documentation says to retry, but it’s not appropriate in the case where you run out of credits. It also shows up when your trial is over.


I have to look at the error_type in the error response and check if it’s “insufficient_quota” and prevent retrying in this case.

Are there better resources for all the different error codes the gpt3 openai endpoint can return?

It’s pretty unclear and there’s very few docs on it.