A company runs a web application that stores user account transaction data in an Amazon RDS for PostgreSQL database and session data in an Amazon DynamoDB table. Users are experiencing slow page load times. A review reveals that the RDS database is overwhelmed with read-only reporting queries, and the DynamoDB table is facing write throttling due to uneven partition distribution. Which two database configurations should a solutions architect implement to resolve these performance issues? (Select TWO.)
- Create Amazon RDS read replicas to handle the read-only reporting queries.Answer
- Use a high-cardinality, randomized attribute as the Amazon DynamoDB partition key to distribute write traffic.Answer
- CUse a monotonically increasing timestamp as the Amazon DynamoDB partition key to store session data in chronological order.
- DConfigure an Amazon RDS read replica to serve as the primary automatic failover target instead of deploying Multi-AZ.
Answer
To resolve the performance issues, the solutions architect should create RDS read replicas to handle the read-only reporting queries, and use a high-cardinality, randomized attribute as the DynamoDB partition key to distribute write traffic.
Creating RDS read replicas allows the reporting queries to be directed away from the primary database, lowering CPU utilization on the writer instance. Implementing a high-cardinality partition key for DynamoDB ensures that write requests are spread evenly across multiple physical partitions, which eliminates write throttling.
Step-by-Step Solution
Key Concept
Scaling read workloads using RDS read replicas and designing DynamoDB partition keys to avoid partition throttling.