AWS Certified SysOps Administrator – AssociateCost and Performance OptimizationHard
A manufacturing company uses an AWS IoT Core solution to ingest telemetry data from thousands of factory sensors. The data is then processed by an AWS Lambda function and stored in Amazon DynamoDB. The SysOps team observes that the Lambda function is occasionally throttling, leading to delayed data processing. The Lambda function processes data in batches and has variable invocation rates due to bursty sensor data. Which action should the SysOps Administrator take to ensure consistent processing performance and optimize costs?
- AConfigure Provisioned Concurrency for the Lambda function.
- BImplement a dead-letter queue (DLQ) for the Lambda function.
- CIncrease the memory allocated to the Lambda function.
- DIncrease the account-level concurrent execution limit for Lambda.
Show answer & explanationAnswer & explanation
Correct answer: A. Configure Provisioned Concurrency for the Lambda function.
Provisioned Concurrency initializes a requested number of execution environments, ensuring that they are ready to respond to invocations with very low latency. This completely eliminates cold starts and reduces throttling during bursty traffic, making it ideal for interactive services or highly variable workloads like bursty IoT data processing where consistent performance is critical.
Why the other options are wrong
- B. A dead-letter queue (DLQ) is for handling failed invocations, not for preventing throttling or improving the performance of successful invocations.
- C. Increasing memory can improve cold start times and execution duration for a single invocation, but it doesn't directly address throttling due to bursty invocation rates or ensure consistent concurrency.
- D. Increasing the account-level limit might allow more concurrent executions, but it doesn't solve the cold start problem or proactively manage concurrency for predictable bursts. Provisioned Concurrency is a more targeted solution for specific functions.
Lambda Provisioned Concurrency
A feature that keeps a specified number of execution environments initialized and ready to respond to invocations, eliminating cold starts.
- Eliminates cold start latency.
- Ensures consistent, low-latency performance.
- Ideal for interactive services and bursty workloads.
Memory trick: Lambda speed: warm up, scale out, or fine-tune code.