An EC2 workload runs in one region on a single instance type. For the last month, CloudWatch metrics show average CPU utilization of 12% and no sustained memory pressure. The team wants to reduce cost while maintaining the current performance level. What is the best first step?

Increase the instance size to reduce the risk of performance regression.

If average CPU is already very low (12%) and there is no sustained memory pressure, increasing the instance size would likely increase cost without providing needed performance benefits. This directly contradicts the evidence of underutilization.

Switch to Spot Instances immediately to reduce cost regardless of utilization.

Switching to Spot could reduce cost, but the question’s evidence indicates a sizing/overprovisioning issue (low utilization). The best first step is to address the primary inefficiency—right sizing—before switching purchase models. Spot also introduces the possibility of interruptions, which is unnecessary if the workload can be optimized through sizing alone.

Disable detailed monitoring to lower CloudWatch charges.

Reducing monitoring cost is usually a minor optimization compared with compute savings. Also, detailed monitoring might still be needed to make accurate rightsizing decisions. The main cost driver indicated by the metrics is idle/overprovisioned compute, which requires sizing analysis rather than lowering monitoring granularity.

What does this SAA-C03 question test?

Read the scenario before looking for a memorised answer.

What is the correct answer to this question?

The correct answer is: Use AWS Compute Optimizer to get recommendations for instance type and size changes. — The best first step is to use AWS Compute Optimizer to perform right sizing recommendations. The metrics show sustained low CPU utilization (12%) with no sustained memory pressure, which strongly suggests the current instance is overprovisioned. Compute Optimizer uses utilization history to recommend a cost-optimized instance type/size that maintains performance targets. The other options either increase cost, introduce unnecessary Spot interruption risk, or optimize monitoring costs that are unlikely to address the largest compute waste. B increases cost despite evidence of underutilization. C addresses purchasing rather than the sizing problem indicated by metrics, and it adds interruption risk that isn’t required to meet the goal. D does not change the compute resources causing the cost inefficiency and focuses on a much smaller cost component.

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.