AI-300 Practice Question 76: APIM Backend Pool Failover
Microsoft Machine Learning Operations Engineer Associate (AI-300) · Test 8 · question 6 of 10
An application uses model deployments in two Azure regions through Azure API Management. When the primary deployment returns 429 or 5xx responses, the team wants requests to fail over to the secondary automatically, and the gateway to stop sending traffic to the unhealthy backend for a period. What should they configure?
Show answer and explanation
Answer:
API Management backends can be grouped in a load-balanced pool, where priority or weight decides which backend is used. A circuit breaker rule on a backend trips after failures such as 429 or 5xx and takes it out of rotation for a set time. Together they provide gateway-level failover (option B).
- A is wrong because Traffic Manager routes by DNS and probes endpoint health. It does not react to 429 responses from individual requests. - C is wrong because this repeats the failover code in every client and has no shared memory of an unhealthy backend. - D is wrong because more quota may reduce throttling, but gives no failover when the primary is unavailable or the region has problems.
Reference: Backends in API Management