fix timeout issue with a pause on retry

[mtmahe]

• [x ] I understand that this repository is auto-generated and my pull request may not be merged

Changes being requested

When running embedding models in azure to create vector indexes there is a timeout even with ADA set to 240 TPM. Simply adding a 10s sleep (only on retry attempts) fixes this.

Thank you so much for all your amazing work!

Additional context & links

[slyapustin]

@mtmahe Any specific reason why it's set to 10 seconds?

[rattrayalex]

@mtmahe – unfortunately this change would add a sleep to all retried requests, not just the ones you are looking to fix, so we cannot accept this change.

cc @kristapratico – sending a retry-after header from the azure API here might be helpful.

[mtmahe]

Hi @rattrayalex

Thanks for consider my request. Yes, a retry-after should work.

[kristapratico]

cc @kristapratico – sending a retry-after header from the azure API here might be helpful.

The Azure embeddings API does return a Retry-After header. There was a bug in the past week where this header wasn't being returned in the response, but it should be restored now. Let us know if the issue persists for you @mtmahe