RDS read traffic has grown while write capacity is sufficient.
Distribute reads instead of enlarging the writer.
Which architecture is most appropriate?
Create an RDS Read Replica and route read traffic to it when read scaling is the limiting requirement.
Detailed explanation
A read replica can absorb read traffic without scaling the primary writer solely for reads.
A read replica can absorb read traffic without scaling the primary writer solely for reads.
This increases write capacity and cost even though writes are not the bottleneck.
This increases write capacity and cost even though writes are not the bottleneck.
CloudFront is not a database read or write scaling feature.
CloudFront is not a database read or write scaling feature.
Stopping reads fails the application requirement.
Stopping reads fails the application requirement.
Check official documentation
Compare the following points with the official documentation.
Amazon RDS Read Replicasの公式説明を確認する。
What to confirm
読み取り負荷と書き込み負荷を分けたDB設計を説明できる。
Key points
- RDS Read Replica
- Read workload
- Primary database
Notes
- Environment: AWS公式RDSドキュメントの確認
- Confirm current AWS service behavior and pricing in the official documentation.
Foundation review
Read scaling
Replicas add read capacity without enlarging the writer just for read traffic.
Workload separation
Separate read and write paths so each can be scaled for its own bottleneck.