A Lambda function behind an API needs consistent low latency. Traffic normally drops to near zero, then spikes several times per hour. During spikes, the p95 latency often spikes above 800 ms due to cold starts. The team wants to keep using Lambda (no containers) but minimize cold start impact during predictable spikes. What is the best AWS configuration to meet this goal?

Turn on AWS CloudTrail data events to monitor cold start frequency and tune the runtime accordingly.

Monitoring does not prevent cold starts; it only provides visibility so it cannot meet the latency requirement by itself.

Configure an ALB with target group health checks to keep Lambda warm by sending periodic requests.

Periodic warm-up calls are not a reliable or scalable mechanism compared to provisioned concurrency for consistent latency.

Increase the function memory size to the maximum and rely on the larger memory to eliminate cold starts.

More memory can improve runtime performance, but it does not guarantee removal of cold starts.

What does this SAA-C03 question test?

Static NAT maps one inside address to one outside address.

What is the correct answer to this question?

The correct answer is: Enable Lambda provisioned concurrency on a published function alias and set the minimum provisioned instances to the baseline expected during spikes. — Provisioned concurrency is designed specifically to reduce or avoid cold start latency by keeping a specified number of Lambda execution environments initialized. The best approach is to publish a version and attach provisioned concurrency to an alias, then set the minimum provisioned instances to cover the expected traffic baseline during predictable spikes. Increasing memory helps compute speed but doesn’t pre-initialize environments. ALB health checks might reduce some cold starts but cannot guarantee consistent low latency for bursty traffic. Monitoring with CloudTrail improves observability, not performance. Why others are wrong: Increasing memory can reduce function duration, but cold start delays still occur when there are no initialized environments. ALB health checks may create artificial warm-up behavior but are not equivalent to Lambda’s deterministic initialization and can fail under sudden bursts or changing traffic patterns. CloudTrail data events provide metrics for troubleshooting, but they do not alter Lambda’s initialization behavior, so they won’t reliably reduce p95 latency during spikes.

What should I do if I get this SAA-C03 question wrong?

Then try more questions from the same exam bank and focus on understanding why the wrong options are tempting.