A Jenkins server on EC2 uses a 500 GB gp2 EBS volume. Nightly builds require sustained 3,000 IOPS, but CloudWatch shows the volume’s BurstBalance hits 0 during builds. Which change will meet the sustained IOPS requirement most cost-effectively?

Changing the EBS volume type from gp2 to gp3 is the most cost-effective solution. gp3 volumes offer a baseline performance of 3,000 IOPS and 125 MiB/s throughput independently of size, which directly meets the requirement without needing to increase storage. gp3 also costs 20% less per GB than gp2. Increasing the gp2 volume to 1,000 GB would provide 3,000 IOPS (3 IOPS/GB 1,000 GB), but it doubles the storage cost unnecessarily. Converting to st1 is unsuitable as it's an HDD designed for throughput, not high random IOPS, and cannot meet the 3,000 IOPS requirement. Switching to io2 would provide the required IOPS but is significantly more expensive than gp3, making it less cost-effective for this workload.