AWS Certified Solutions Architect - Associate

Explanation

The correct answer is D because:

1. Predictable Peak Hours: The problem states that application peak hours occur at the same time each day, which is a predictable pattern.

2. Performance Issue Timing: Users report slow performance at the start of peak hours, but the application performs normally 2-3 hours after peak hours begin. This indicates that the Auto Scaling group eventually scales up based on metrics (likely CPU utilization), but this reactive scaling happens too slowly.

3. Scheduled Scaling vs. Dynamic Scaling:

   • Dynamic scaling (options B and C) reacts to metrics like CPU or memory utilization. This is reactive and takes time to trigger and complete (instance launch time + application warm-up time).
   • Scheduled scaling (option D) proactively scales capacity before the predictable peak hours begin, ensuring sufficient capacity is available when needed.

4. Why not the other options:

   • A: An Application Load Balancer helps distribute traffic but doesn't address the capacity issue - it would just distribute traffic to insufficient instances.
   • B & C: Dynamic scaling based on memory or CPU utilization would still cause the initial performance degradation because scaling would only trigger AFTER metrics exceed thresholds, and there's a delay for instances to launch and become ready.

5. Best Practice: For predictable traffic patterns, AWS recommends using scheduled scaling actions to ensure capacity is available before the load increases, rather than waiting for metrics to trigger scaling.

Key AWS Concepts:

• Scheduled Scaling: Best for predictable, recurring patterns
• Dynamic Scaling: Best for unpredictable, variable patterns
• Instance Launch Time: EC2 instances take time to launch (minutes) and applications may need additional warm-up time