Containers in multiple Availability Zones share an EFS file system. Load is low at night, but read and write throughput spikes whenever the morning job starts.
The team cannot forecast the peak accurately and wants to avoid fixed provisioned charges while preventing throughput constraints during spikes.
Which EFS configuration is most appropriate?
Use EFS Elastic throughput for unpredictable spikes and monitor metrics such as MeteredIOBytes to manage usage, performance, and cost.
Detailed explanation
This removes shared-file and Availability Zone resilience requirements and does not solve the throughput-mode decision.
This removes shared-file and Availability Zone resilience requirements and does not solve the throughput-mode decision.
Elastic throughput automatically scales with workload activity and is suited to unpredictable spikes. Monitoring provides visibility into usage and cost.
Elastic throughput automatically scales with workload activity and is suited to unpredictable spikes. Monitoring provides visibility into usage and cost.
Provisioned throughput fits a known performance requirement, but an unpredictable peak makes fixed capacity and its charge inefficient here.
Provisioned throughput fits a known performance requirement, but an unpredictable peak makes fixed capacity and its charge inefficient here.
Bursting throughput is tied to stored data, so a small file system or sustained load can become throughput-constrained.
Bursting throughput is tied to stored data, so a small file system or sustained load can become throughput-constrained.
Check official documentation
Compare the following points with the official documentation.
Amazon EFS performance specificationsとCloudWatch metricsの公式説明を確認する。
Official references
What to confirm
Elastic、Provisioned、Burstingの選択基準を、負荷の予測可能性と課金・性能の関係から説明できる。
Key points
- Elastic throughput
- Provisioned throughput
- Bursting throughput
- MeteredIOBytes
- Throughput
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 the behavior in the official SOA-C03 guide and Amazon EFS documentation.
Foundation review
Predictability
Elastic throughput is suited to spiky or difficult-to-forecast workloads, while Provisioned throughput fits a known requirement.
Operational guardrail
Monitoring throughput usage helps detect unexpected demand and keep performance and cost within the plan.