Back to Amazon RDS

Amazon RDS Provisioned IOPS

5 minutes 5 Questions

Amazon RDS Provisioned IOPS (Input/Output Operations per Second) is designed to meet the needs of heavy and demanding applications with high transactional workloads. It allows you to provision up to 64,000 IOPS per DB instance, providing predictable and consistent performance. With Provisioned IOPS…

Test mode:
Start practice test
AWS Certified Solutions Architect - Amazon RDS Provisioned IOPS Example Questions

Test your knowledge of Amazon RDS Provisioned IOPS

Question 1

An Amazon RDS MySQL instance for a highly transactional OLTP workload shows low CPU utilization and normal buffer cache performance. CloudWatch indicates ReadIOPS/WriteIOPS are consistently at the maximum for its gp2 storage, with elevated disk queue depth, while the instance’s EBS bandwidth is not saturated. The workload is write‑heavy, so offloading reads will not help much. You need sustained higher storage IOPS with predictable latency. What should you do to meet the performance requirement?

Correct Answer: Increase the IOPS by moving to Provisioned IOPS

Provisioned IOPS storage for Amazon RDS is designed to deliver consistent, high I/O performance for I/O‑intensive OLTP workloads. When CloudWatch shows storage IOPS saturation on gp2 and CPU/memory and EBS bandwidth are not the bottleneck, switching the storage type to Provisioned IOPS and provisioning the required IOPS provides predictable latency and higher sustained IOPS. Multi‑AZ improves availability, not throughput, and can increase write latency. Read replicas primarily offload reads and do not help a write‑heavy workload. Increasing the instance size does not resolve a storage‑level IOPS cap when compute and bandwidth are not constrained.

Question 2

During peak traffic, an Amazon RDS MySQL instance using General Purpose SSD (gp2) shows WriteLatency > 20 ms and spikes in VolumeQueueLength, while CPU utilization remains below 40% and memory is not constrained. The application is write-heavy, and user requests slow down during these periods. What is the best action to address the performance bottleneck?

Correct Answer: Configure RDS for Provisioned IOPS and increase IOPS capacity

CloudWatch indicates the workload is I/O-bound on writes (high write latency and queue depth) rather than CPU or memory constrained. Moving the RDS instance to Provisioned IOPS (io1/io2) storage and allocating sufficient IOPS directly addresses the bottleneck by providing predictable, higher I/O throughput and lower latency for write-heavy workloads. Adding a read replica does not help because replicas scale reads, not writes. Vertically scaling CPU and memory does not resolve storage I/O contention. Migrating to Amazon Redshift is inappropriate for OLTP workloads and will not solve low-latency transactional write performance issues.

Question 3

Your RDS instance is not meeting the desired IOPS required for a particular business-critical transaction. The instance is already on Provisioned IOPS storage. What would be the best solution to increase IOPS with minimal downtime?

Correct Answer: Increase the allocated storage size to automatically scale the Provisioned IOPS

Increasing the allocated storage size automatically scales the Provisioned IOPS, providing a minimal downtime solution. Converting to General Purpose storage, rebooting the instance, creating a read replica, enabling Multi-AZ deployment or creating a new RDS instance would not help in increasing the IOPS without significant downtime.

More Amazon RDS Provisioned IOPS questions
15 questions (total)
Start 15 question test