Answer: Increase the number of IOPS on the gp3 volume.

## Explanation **Correct Answer: B - Increase the number of IOPS on the gp3 volume.** **Why this is correct:** 1. **gp3 volumes have configurable IOPS**: Unlike gp2 volumes where IOPS are tied to volume size, gp3 volumes allow you to provision IOPS independently of storage size. 2. **Current limitation**: The problem states that performance degrades when IOPS exceed 20,000. With a 2,000 GB gp3 volume, the baseline performance is 6,000 IOPS (3 IOPS per GB), but gp3 allows you to provision additional IOPS up to 16,000 IOPS (for a total of 22,000 IOPS) at no additional cost for the first 3,000 provisioned IOPS. 3. **Cost-effective solution**: Increasing IOPS on a gp3 volume is more cost-effective than migrating to io2 volumes, which are designed for mission-critical applications requiring higher durability and performance guarantees. 4. **Simple scaling**: You can increase IOPS on gp3 volumes without downtime or data migration. **Why other options are incorrect:** **A. Replace the volume with a magnetic volume.** - Magnetic volumes have very low performance (40-200 IOPS) and are not suitable for database workloads. - This would severely degrade performance rather than improve it. **C. Replace the volume with a Provisioned IOPS SSD (io2) volume.** - While io2 volumes offer higher performance (up to 256,000 IOPS) and better durability, they are more expensive. - The problem can be solved more cost-effectively by simply increasing IOPS on the existing gp3 volume. - io2 is typically used for mission-critical databases requiring the highest performance and durability. **D. Replace the 2,000 GB gp3 volume with two 1,000 GB gp3 volumes.** - This doesn't solve the IOPS limitation problem. - Two 1,000 GB gp3 volumes would each have 3,000 baseline IOPS (total 6,000), which is actually worse than the current 2,000 GB volume with 6,000 baseline IOPS. - Splitting volumes doesn't aggregate IOPS for RDS - the database would still be limited by the IOPS of the primary volume. **Key AWS Concepts:** - **gp3 volumes**: Provide baseline performance of 3,000 IOPS for volumes up to 1 TB, plus 3 IOPS per GB beyond 1 TB. - **IOPS scaling**: gp3 allows you to provision additional IOPS independently of storage size. - **Performance monitoring**: Use Amazon CloudWatch metrics like `ReadIOPS`, `WriteIOPS`, and `VolumeQueueLength` to monitor database performance. - **RDS storage optimization**: Always monitor database performance metrics and adjust storage configuration based on actual workload patterns.

Author: LeetQuiz Editorial Team