Several EC2 instances appear oversized compared with their CPU and memory utilization.
The team wants to narrow the candidates after comparing performance risk and price impact.
Which investigation method is most appropriate?
Use AWS Compute Optimizer to analyze utilization and compare instance recommendations, performance risk, and price impact, then narrow the scope with tags.
Detailed explanation
CloudTrail audits API activity; it does not provide continuous resource utilization or right-sizing recommendations.
CloudTrail audits API activity; it does not provide continuous resource utilization or right-sizing recommendations.
A uniform reduction ignores workload requirements and can cause performance or availability problems.
A uniform reduction ignores workload requirements and can cause performance or availability problems.
Compute Optimizer analyzes historical utilization, presents EC2 size recommendations with performance risk and price differences, and supports filtering by tags and Regions.
Compute Optimizer analyzes historical utilization, presents EC2 size recommendations with performance risk and price differences, and supports filtering by tags and Regions.
Security groups control network permissions, not instance size or resource utilization.
Security groups control network permissions, not instance size or resource utilization.
Check official documentation
Compare the following points with the official documentation.
AWS Compute OptimizerのEC2推奨事項、利用率グラフ、タグフィルターの公式説明を確認する。
Official references
What to confirm
一律変更ではなく、利用状況・性能リスク・料金を比較してEC2のサイズ候補を評価できる。
Key points
- AWS Compute Optimizer
- Rightsizing
- Performance risk
- Utilization metrics
Notes
- Environment: AWS公式SOA-C03試験ガイドとAWS公式ドキュメントの確認
- Confirm current behavior in the official SOA-C03 guide and AWS service documentation.
Foundation review
Evidence before change
Recommendations based on observed utilization are safer than applying the same size change to every workload.
Performance and cost
A lower price is useful only when the recommended capacity remains within the workload's performance requirements.