AWS Certified DevOps Engineer – ProfessionalResilient Cloud SolutionsMedium

A global e-commerce company experiences seasonal traffic spikes, particularly during holiday sales events. Their current application runs on Amazon EC2 instances within an Auto Scaling group, fronted by an Application Load Balancer (ALB). During peak times, customers report slow loading pages and occasional service unavailability, even though the Auto Scaling group is configured to scale out. The operations team observes that the EC2 instances are often CPU-bound before new instances become available to handle the increased load. Which strategy should the company implement to proactively address the performance degradation during anticipated traffic surges?

  1. ASwitch to Amazon EC2 Spot Instances for all application servers to reduce costs and allow for more instances to be provisioned.
  2. BIncrease the maximum capacity of the Auto Scaling group and manually scale out instances several hours before the anticipated peak.
  3. CImplement scheduled scaling actions for the Auto Scaling group to increase desired capacity ahead of known peak events.
  4. DConfigure a target tracking scaling policy based on network I/O utilization instead of CPU utilization.
Show answer & explanation

Correct answer: C. Implement scheduled scaling actions for the Auto Scaling group to increase desired capacity ahead of known peak events.

Scheduled scaling actions allow an Auto Scaling group to proactively adjust its capacity based on predictable demand patterns, such as seasonal sales. This ensures that new instances are provisioned and ready before the traffic surge hits, preventing performance degradation caused by reactive scaling delays.

Why the other options are wrong

  • A. Spot Instances are not suitable for critical, uninterrupted workloads due to their ephemeral nature and potential for interruption, which could worsen availability during peak times.
  • B. Manually scaling out is reactive and prone to human error; scheduled actions automate this process reliably.
  • D. Changing the metric might help, but it doesn't address the core problem of reactive scaling not being fast enough for predictable spikes. Proactive scaling is needed.

Scheduled Scaling

A feature of AWS Auto Scaling that allows you to adjust the capacity of your Auto Scaling group based on a predictable schedule.

  • Ideal for predictable traffic patterns (e.g., daily, weekly, seasonal spikes).
  • Ensures resources are provisioned proactively before demand increases.
  • Can be used to scale both out and in.

Memory trick: Schedule your scaling, save your servers from suffering.

More Resilient Cloud Solutions questions