SAA-C03 Practice Question 37: Aurora Reader Endpoint Scaling

Aurora Replicas share the cluster's storage volume with the writer, so they add read capacity without copying data, and they can be added or removed as demand changes. The reader endpoint is a single DNS name that load-balances new connections across the available replicas, so the application holds one endpoint and gains capacity automatically as replicas are added. Reporting queries then run on instances separate from the writer, which relieves the contention with transactional writes.

Option A solves a different problem at much greater cost. Aurora Global Database targets cross-Region disaster recovery and low-latency reads for distant users; it is heavy machinery for read scaling inside one Region.

Option C buys time rather than separating the workloads. A bigger writer still runs reporting and transactions on the same instance, so a heavy analytical query still competes with writes, and the ceiling returns as usage grows.

Option D can help for repeated identical lookups, but it requires application changes to manage the cache and does not suit varied analytical queries, whose result sets are large and rarely repeated.

Reference: Amazon Aurora connection management and endpoints