SOA-C02 Cost and Performance Optimization Practice Question

A company runs a web application on EC2 instances in an Auto Scaling group. The application experiences variable traffic. The company wants to improve performance and reduce costs. Which THREE actions should the company take?

⚠ Common exam trap

SOA-C02 often tests the misconception that cost optimization means simply reducing instance count or resizing instances, when the correct answer is almost always elasticity (dynamic scaling) plus purchase-option mixing.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

Implement dynamic scaling policies based on CPU utilization.

Option A is correct because dynamic scaling policies based on CPU utilization let the Auto Scaling group add instances when demand rises and remove them when demand falls, directly improving performance during peaks while reducing cost during troughs. Option C is correct because an Application Load Balancer distributes incoming HTTP/HTTPS traffic across healthy instances and connection draining (deregistration delay) allows in-flight requests to complete before an instance is terminated during scale-in, preventing errors and improving availability. Option D is correct because mixing On-Demand and Spot Instances in the Auto Scaling group lowers compute costs by using discounted Spot capacity for fault-tolerant portions of the workload while On-Demand instances provide baseline reliability. Option B is not appropriate because simply reducing the number of instances lowers capacity and would degrade performance under variable traffic rather than improve it. Option E is not appropriate because increasing instance size (vertical scaling) is less elastic and typically more expensive than horizontal scaling, and it does not efficiently match variable traffic or reduce costs.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✓
Implement dynamic scaling policies based on CPU utilization.
Why this is correct
Dynamic scaling policies using a target tracking policy based on CPU utilization are correct because they automatically adjust the Auto Scaling group's desired capacity in real time to maintain CPU at a defined target (e.g., 60%). This prevents both over-provisioning and under-provisioning, enabling the web application to handle variable traffic without manual intervention, and it optimizes cost by only adding instances when demand actually increases.
✗
Reduce the number of instances in the Auto Scaling group to lower costs.
Why it's wrong here
Reducing the number of instances in the Auto Scaling group is incorrect because it lowers capacity across the board, which will likely cause performance degradation and potential request timeouts or throttling during traffic spikes. This manual action sacrifices availability to achieve cost savings, whereas a more appropriate approach is to use dynamic scaling to keep only the necessary capacity and scale out precisely when needed.
✓
Use an Application Load Balancer with connection draining.
Why this is correct
Using an Application Load Balancer with connection draining is a valid complementary practice because it ensures that in-flight requests are completed before an instance is terminated during a scale-in event. This prevents user-facing errors and improves reliability during elastic scaling, as the ALB stops sending new connections to an instance in the draining state while allowing existing sessions to finish cleanly.
✓
Use a mix of On-Demand and Spot Instances in the Auto Scaling group.
Why this is correct
Using a mix of On-Demand and Spot Instances in the Auto Scaling group is a correct cost-optimization method that maintains reliability. On-Demand instances provide baseline capacity for critical workloads, while Spot Instances handle flexible or batch processing at substantial discounts; however, because Spot capacity can be interrupted, you should design the mixture so that the workload can tolerate Spot reclamation, for example by using instance weighting and multiple instance types.
✗
Increase the instance size to handle peak load.
Why it's wrong here
Increasing the instance size to handle peak load is incorrect because it relies on vertical scaling, which is expensive and still leaves a single point of failure; the larger instance also remains idle during low-traffic periods, wasting money. Horizontal scaling via Auto Scaling with multiple smaller instances is more elastic and fault-tolerant, and combining it with dynamic scaling policies provides a cost-effective solution that matches capacity to actual demand.

Go deeper

SOA-C02 Cost and Performance Optimization Practice Question

Why each option matters

Option-by-option breakdown

Related to this question