AWS Certified SysOps Administrator – AssociateCost and Performance OptimizationEasy
A company is using an Amazon RDS for PostgreSQL database instance that is experiencing high CPU utilization during peak hours. The current instance type is db.m5.large. The database load is primarily read-heavy. A SysOps Administrator needs to optimize performance without significantly increasing costs or downtime for write operations. What is the most effective solution?
- AUpgrade the RDS instance to a larger instance type (e.g., db.m5.xlarge).
- BMigrate the database to Amazon Aurora PostgreSQL.
- CEnable Multi-AZ deployment for the RDS instance.
- DCreate an Amazon RDS Read Replica and offload read queries to it.
Show answer & explanationAnswer & explanation
Correct answer: D. Create an Amazon RDS Read Replica and offload read queries to it.
For read-heavy workloads, creating an RDS Read Replica allows you to offload read traffic from the primary instance, thereby reducing its CPU utilization and improving overall performance without impacting write operations. This is a cost-effective way to scale read capacity.
Why the other options are wrong
- A. Upgrading the instance type would increase costs significantly and might involve downtime. It's a valid option for general performance but not the most effective for read-heavy workloads specifically.
- B. Migrating to Aurora PostgreSQL would offer significant performance benefits, but it's a more complex and potentially more expensive solution than simply adding a Read Replica for a read-heavy workload.
- C. Multi-AZ improves availability and durability but does not directly address high CPU utilization from a read-heavy workload, as all reads typically go to the primary instance unless specifically configured otherwise (which is not the primary purpose of Multi-AZ).
RDS Read Replicas
Copies of your primary RDS database instance that are used to offload read traffic and improve read scalability.
- Asynchronous replication.
- Improves read performance and scalability.
- Can be promoted to a standalone database if needed.
Memory trick: Making RDS faster: scale up, spread out, or specialize reads.