AZ400-Q: set a retry policy
A deployment task occasionally fails because a cloud control-plane API returns a documented transient throttling response. What retry policy is appropriate?
- A) Retry every failure forever.
- B) Use bounded retries with backoff for the understood transient condition, log attempts, and fail/escalate when the limit is reached. ✓
- C) Treat all failures as success.
- D) Remove timeouts.
Answer: B. Bounded, observable retries handle known transient faults without hiding persistent failures or consuming unlimited capacity.
Why not the others: Infinite/silent retries mask defects; removing limits increases runaway risk.