Step Scaling

Correct Answer: PercentChangeInCapacity

PercentChangeInCapacity adjusts the Auto Scaling group by a percentage of its current size, making each step proportional to existing capacity. This is ideal when traffic spikes scale relative to current load (e.g., doubling or tripling), because larger groups will add more instances in absolute terms than smaller groups. ChangeInCapacity uses a fixed instance count per step, which does not scale proportionally and can be too small for large groups or too large for small groups. ExactCapacity sets the group to a specific size, which is not suitable for dynamic, proportional responses to varying spikes. AbsoluteCapacity is not a valid step scaling adjustment type.

Correct Answer: Define distinct CloudWatch alarm thresholds for each step.

Defining distinct CloudWatch alarm thresholds for different steps ensures that the scaling policy reacts accurately to different levels of workload. Creating one alarm, setting static instance counts, and manually setting alarms are not suitable for varying requirements. Defining only one step adjustment or using custom metrics based on instance types may be insufficient.

Correct Answer: Create a step scaling policy with multiple steps, specifying different scaling adjustments for different alarm thresholds.

Step scaling policies allow multiple steps with different scaling adjustments based on the alarm thresholds, providing a better response to sudden traffic changes. Simple scaling and target tracking policies are not as suitable to handle rapid changes in traffic. Manual scaling is inefficient and AWS Instance Scheduler and Trusted Advisor do not serve this purpose.