AWS Certified CloudOps Engineer Associate SOA-C03 Practice Question

An online learning platform runs its stateless web tier in an Auto Scaling group of Amazon EC2 instances behind an Application Load Balancer (ALB). During live events, incoming traffic can surge from a few hundred to tens of thousands of requests per minute within a short time. The group currently uses a simple scaling policy that adds one instance when average CPUUtilization exceeds 60% for 5 minutes, but users still experience high latency during sudden spikes. The platform needs a solution that reacts more quickly to large increases in load while still allowing the fleet to scale in smoothly when demand subsides. Which change will best meet these requirements?

Enable predictive scaling for the Auto Scaling group using CPUUtilization as the forecast metric to add capacity ahead of traffic spikes.
Replace the simple scaling policy with a target-tracking policy that maintains the ALB RequestCountPerTarget metric at a chosen value (for example, 100 requests per target).
Add a scheduled scaling action to increase desired capacity to the maximum group size shortly before each live event.
Reduce the cooldown period and modify the simple scaling policy to launch three instances when CPUUtilization exceeds 60% for 5 minutes.

Report Issue

Answer Description

Target-tracking scaling policies continually adjust desired capacity to keep a specified metric, such as ALB RequestCountPerTarget, close to the chosen target value. When requests surge, the policy increases capacity by more than one instance if necessary, allowing the fleet to react faster than a simple scaling policy that always adds a fixed number of instances. As load drops, the same policy gradually scales the group back in, without the need for manually maintained threshold tiers. Decreasing cooldowns or increasing step size in a simple policy may still under- or over-react and require tuning, scheduled scaling cannot handle unpredictable spikes, and predictive scaling based on CPU will not be effective when traffic patterns are irregular or when request-based metrics are a better indicator of load.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.