AWS Certified SysOps Administrator – AssociateCost and Performance OptimizationEasy
A company is using Amazon EC2 instances to host a stateless web application. The application's traffic fluctuates significantly throughout the day, with peak usage during business hours and very low usage overnight. The SysOps Administrator wants to minimize EC2 costs while ensuring the application can handle peak loads and automatically scale down during off-peak times. Which combination of EC2 features should be implemented?
- AReserved Instances with scheduled scaling.
- BAuto Scaling Group with dynamic scaling policies and Spot Instances.
- CAuto Scaling Group with dynamic scaling policies and On-Demand Instances.
- DOn-Demand Instances with a fixed number of instances.
Show answer & explanationAnswer & explanation
Correct answer: C. Auto Scaling Group with dynamic scaling policies and On-Demand Instances.
An Auto Scaling Group combined with dynamic scaling policies (e.g., target tracking) allows the number of EC2 instances to automatically adjust based on real-time demand, ensuring capacity for peak loads and scaling down during off-peak times to save costs. On-Demand instances are suitable for the baseline and fluctuating capacity, as the application is critical and stateless.
Why the other options are wrong
- A. Reserved Instances are for predictable, consistent usage and don't provide the flexibility to scale down completely to zero or near-zero, which is needed for highly fluctuating workloads. Scheduled scaling can help but doesn't react dynamically to real-time fluctuations.
- B. Spot Instances are generally not suitable for critical, stateless web applications that cannot tolerate interruptions. While cost-effective, they introduce availability risks that might not be acceptable for a web application.
- D. A fixed number of On-Demand instances would either be over-provisioned during off-peak hours (high cost) or under-provisioned during peak hours (poor performance).
Auto Scaling Group with Dynamic Scaling
An EC2 feature that automatically adjusts the number of instances in response to real-time demand, optimizing performance and cost.
- Scales compute capacity up and down.
- Uses metrics (e.g., CPU utilization) for scaling decisions.
- Ensures high availability and fault tolerance.
Memory trick: Ride the EC2 waves: auto-scale and pay for what you need.