The best approach is Configure a scheduled scaling action to increase the desired capacity to 20 shortly before the business day begins. Since the companys traffic pattern is predictable and starts at the same time every day, scheduled scaling is the most cost-effective and performance-oriented solution. By pre-scaling the infrastructure, the system can handle the morning workload without waiting for metrics like CPU usage to trigger a response. This ensures staff have a responsive application experience right from the start of their day.

The option Set up step scaling with a lower CPU threshold and shorten the cooldown interval for faster instance launches relies on reactive scaling. While it may reduce the delay in scaling, it still waits for load metrics to reach a threshold, which means users may still experience slowness before the scaling kicks in.

The choice Use target tracking scaling with a reduced CPU target threshold and a shorter cooldown period has similar drawbacks. It depends on live metric changes and doesn't guarantee readiness ahead of time. Even with faster reactions, the system still wont have the full capacity ready before peak demand starts.

The option Schedule a scaling action that fixes both minimum and maximum instance counts at 20 just before the workday starts would solve the performance issue, but its unnecessarily costly. By setting both min and max to 20, the Auto Scaling group is prevented from scaling down when demand drops, wasting resources after peak hours.

This question is useful for evaluating how well you understand scaling strategies, especially in the context of predictable workloads. A helpful trick is to look for time-based patterns. When traffic spikes are predictable, scheduled scaling is almost always more efficient and cost-effective than metric-based scaling.

The difficulty in this question lies in differentiating between proactive and reactive scaling strategies. While all the choices relate to Auto Scaling, only scheduled scaling provides guaranteed readiness at a known time with minimal resource waste.

Full AWS Practitioner Certification Question