AI-103 Certification Practice Question #39

Question

You have a customer support agent built by using the Microsoft Foundry Agent Service. The agent calls an Azure OpenAI model deployment.
During load testing, calls intermittently fail and return an HTTP 429 rate limit exceeded error.
You need to handle throttling to reduce call failures and improve reliability under load. The solution must remain within the service and model limits.
What should you do?

Accepted Answer

An HTTP 429 means the Azure OpenAI deployment's rate limit (TPM/RPM) was exceeded. MS Learn's quota guidance prescribes handling this by automatically retrying with **exponential backoff plus random jitter** (honoring `retry-after-ms` when present), which spaces out and randomizes retries so the workload stays within service and model limits and recovers gracefully under load. Retrying immediately on a new thread (A) keeps pressure on the throttled endpoint and makes throttling worse — MS Learn explicitly warns that resending without backing off aggravates the limit. Reducing registered tools (B) doesn't change the rate limit, and splitting uploads into smaller files (D) addresses payload size rather than request throttling. The Suggested Answer (C) and MS Learn agree directly, giving high confidence.

More AI-103 practice questions