Question #659
A solutions architect must evaluate a company’s Amazon EC2 instances and Amazon Elastic Block Store (EBS) volumes to optimize costs for a web application experiencing unpredictable traffic patterns. The application runs on multiple large, general-purpose EC2 instances, and the company has not identified consistent utilization trends. The architect needs to analyze the environment and implement cost-effective improvements. Which solution BEST meets these requirements?
Use AWS Systems Manager OpsCenter to create a dashboard visualizing Amazon CloudWatch metrics for the EC2 instances and EBS volumes. Periodically review the dashboard to identify usage trends, then manually rightsize instances based on observed peaks.
Enable Amazon CloudWatch detailed monitoring for the EC2 instances and EBS volumes. Build a custom dashboard to analyze metrics, manually identify patterns, and resize instances according to peak utilization metrics.
Install the Amazon CloudWatch agent on all EC2 instances. Activate AWS Compute Optimizer and allow it to collect data for at least 12 hours. Rightsize instances based on Compute Optimizer\u2019s automated recommendations.
Upgrade to the AWS Enterprise Support plan. Enable AWS Trusted Advisor and wait 12 hours for analysis. Implement Trusted Advisor\u2019s recommendations to rightsize EC2 instances.
Explanation
Answer C is correct because:
1. AWS Compute Optimizer uses machine learning to analyze historical EC2 and EBS usage, providing automated rightsizing recommendations. This is ideal for unpredictable traffic, as manual analysis (Options A/B) may miss patterns.
2. Installing the CloudWatch agent ensures detailed metric collection, enabling Compute Optimizer to generate accurate insights.
3. Compute Optimizer requires 12+ hours of data to produce recommendations, aligning with the question's context.
Why other options are incorrect:
- A/B: Manual dashboard reviews are error-prone and inefficient for volatile workloads.
- D: Trusted Advisor offers high-level cost checks but lacks Compute Optimizer's granular, ML-driven analysis. Upgrading to Enterprise Support is unnecessary if Compute Optimizer suffices.
Key Points:
- Use Compute Optimizer for automated rightsizing with ML.
- Collect sufficient data (≥12 hours) for accurate recommendations.
- Avoid manual methods for unpredictable traffic patterns.
Answer
The correct answer is: C