Aurora – Custom Endpoints

• Aurora allows you to group a subset of DB instances within a cluster and expose them through a Custom Endpoint.

• This is especially useful for workload isolation, such as running analytical queries or batch jobs on specific replicas without impacting other replicas or the writer.

• Once a Custom Endpoint is configured, the Reader Endpoint can still be used for general-purpose reads, but applications requiring targeted workloads should use the Custom Endpoint instead.

Example Use Case

• Writer Endpoint – directs all write operations to the primary instance.

• Reader Endpoint – balances general read traffic across all read replicas.

• Custom Endpoint – routes queries to selected replicas (e.g., larger instance classes like db.r5.2xlarge) dedicated to analytics.

Benefits

• Query isolation – separates heavy analytical workloads from regular reads, preventing performance degradation.

• Instance targeting – choose replicas optimized for a specific workload or hardware profile.

• Improved resource utilization – ensures that high-CPU or high-memory queries don’t interfere with transactional workloads.