Microsoft Certified: Azure AI Engineer AssociatePlan and manage an Azure AI solutionMedium
A research institution is developing an Azure AI solution to analyze large volumes of scientific literature for patterns and insights. The solution uses Azure Cognitive Search to index the documents and provide search capabilities. They anticipate that the indexing process will be computationally intensive and may occasionally exceed the default request limits of a standard Cognitive Search resource, leading to throttling. To avoid service disruptions and ensure consistent performance during peak indexing, what is the most cost-effective way to handle these intermittent bursts of high indexing demand?
- AUpgrade the Cognitive Search service to a higher tier permanently.
- BImplement a retry logic with exponential backoff in the indexing client.
- CScale out the Cognitive Search service by adding more replicas.
- DTemporarily scale up the Cognitive Search service by increasing partitions during peak times.
Show answer & explanationAnswer & explanation
Correct answer: B. Implement a retry logic with exponential backoff in the indexing client.
Implementing retry logic with exponential backoff is a highly cost-effective and robust strategy for handling intermittent throttling. It allows the client to automatically reattempt requests after increasing delays, gracefully handling temporary service limitations without incurring additional costs for scaling up or out when not consistently needed.
Why the other options are wrong
- A. Upgrading to a higher tier permanently is expensive if the peak demand is only intermittent and not continuous.
- C. Scaling out by adding replicas is for increasing query throughput and availability, not primarily for overcoming intermittent indexing throttling. It also increases cost.
- D. Temporarily scaling up by increasing partitions is an option to increase indexing capacity, but it incurs immediate additional costs for the duration and requires manual intervention or automation, which can be more complex and costly than client-side retry logic for intermittent issues.
Retry Logic with Exponential Backoff
A client-side pattern for handling transient errors (like throttling) by retrying failed operations with progressively longer delays, preventing overwhelming the service.
- Gracefully handles intermittent service limitations.
- Prevents overwhelming the service with repeated requests.
- Cost-effective as it avoids unnecessary scaling of the backend service.
Memory trick: Backoff 'Retries' save costs on temporary throttling.