Examkin

Question

Difficulty: MediumOptimizing Compute and Storage Performance

An e-commerce company operates a high-traffic catalog search application. The application's backend database runs on an Amazon RDS for PostgreSQL DB instance configured in a Multi-AZ deployment with 3,000 Provisioned IOPS (SSD) storage. During peak shopping events, read latency on the database increases significantly, and CPU utilization on the DB instance exceeds 90%90\%, leading to slower search response times. The write volume remains low and stable. Which action should a solutions architect take to resolve the database performance bottleneck with minimal latency?

  1. Deploy Amazon RDS Read Replicas in the active Availability Zones, and update the application configuration to route all read-only search queries to the read replica endpoints.Answer
  2. B
    Configure the application to route read-only search queries to the endpoint of the existing Multi-AZ standby DB instance to balance the query load.
  3. C
    Request AWS Support to pre-warm the database's front-end Application Load Balancer to handle the sudden increase in connection requests during peak events.
  4. D
    Decrease the scale-out cooldown period on the database's Auto Scaling group to launch additional DB instances faster when CPU utilization exceeds 90%90\%.

Answer

Deploy Amazon RDS Read Replicas in the active Availability Zones, and update the application configuration to route all read-only search queries to the read replica endpoints.
Deploying Amazon RDS Read Replicas is the correct action because it directly offloads read-heavy queries from the primary DB instance. This reduces both the CPU load and Provisioned IOPS consumption on the primary database, resolving the catalog search latency bottleneck. Read replicas are designed exactly for scaling read-heavy workloads horizontally.

Step-by-Step Solution

1
Analyze the metrics and workload patterns.
Identify that the database has high CPU (>90%>90\%) and high read latency during peak times, while write traffic remains low and stable.
This confirms that the bottleneck is read-heavy compute and storage operations, making the workload a prime candidate for read scaling.
2
Evaluate the current architecture capabilities.
The current Multi-AZ standby instance is passive and cannot serve read traffic.
Recognizing that Multi-AZ is for durability and failover, not read scaling, prevents incorrect routing configurations to the standby.
3
Select the correct horizontal scaling mechanism.
Deploy one or more RDS Read Replicas and point the application's search queries to the replica endpoints.
Read replicas run active database engines that replicate data asynchronously, allowing them to offload read operations from the primary instance, thus reducing CPU and storage utilization on the primary.

Key Concept

Scaling read performance in Amazon RDS using Read Replicas vs Multi-AZ deployments
Rate this question

Topics

Optimizing Compute and Storage Performance