A provisioned Aurora MySQL cluster has a writer and at least one reader, all available. Adjust reader count between one and five toward average reader CPU of 50% as read demand changes.
The application sends writes to the writer and new read connections to the reader endpoint, supports newly added readers, and has the required scaling permissions and cooldown configured. Automatic redistribution of every individual query is not required.
Which configuration is most appropriate?
Use Aurora Auto Scaling to adjust reader count within capacity bounds and the reader endpoint to distribute new read connections.
Detailed explanation
Without readers, the average reader metric cannot be evaluated. Maintain at least one reader for this policy.
Without readers, the average reader metric cannot be evaluated. Maintain at least one reader for this policy.
The endpoint balances connections; it does not create or delete readers in place of a scaling policy.
The endpoint balances connections; it does not create or delete readers in place of a scaling policy.
This feature adjusts reader count rather than writer size and does not implement the average-reader-CPU requirement.
This feature adjusts reader count rather than writer size and does not implement the average-reader-CPU requirement.
It adjusts reader count and the reader endpoint distributes new connections; this does not automatically resize the writer.
It adjusts reader count and the reader endpoint distributes new connections; this does not automatically resize the writer.
Check official documentation
Compare the following points with the official documentation.
Review the linked AWS documentation for storage, connection paths, scaling scope, and prerequisites.
Official references
What to confirm
Explain the operational limits of the alternatives and the application checks needed after configuration.
Key points
- Aurora Auto Scaling
- Aurora Replicas
- Reader endpoint
Notes
- Environment: AWS SOA-C03 exam guide and the linked AWS documentation
- Reviewed against the linked AWS documentation on 2026-10-06.
Foundation review
Scaling scope and prerequisites
Aurora Auto Scaling adjusts readers, not writer load. Average reader CPU includes manually created readers and excludes the writer. Check at least one reader, an available cluster, capacity bounds, scale-in, and cooldown. It removes only readers it created, not manually created readers. Provisioning takes time and the CPU target is not guaranteed continuously.
The reader endpoint balances connections
The reader endpoint balances new connections among available readers, not each query within an existing connection. Manage long-lived connections and pools so new readers can receive connections. With no readers it connects to the writer; the endpoint name alone does not guarantee a read-only session.