Boosting App Resiliency: Migrating to High-Performance GP3 Volumes

Traditionally, if your PostgreSQL database or high-throughput application needed more than 16,000 IOPS on AWS, you were forced into a corner. You either had to pay a premium for io2 Block Express volumes or manage a complex, risky array of striped (RAID 0) volumes.

With the latest updates to Amazon EBS gp3 volumes, that tradeoff is largely gone. By supporting up to 64 TiB of storage, 80,000 IOPS, and 2,000 MiB/s throughput on a single volume, you can now simplify your architecture significantly. In this guide, we'll help FinCore Systems—a fintech startup—migrate their high-frequency trading logs from a fragile RAID-0 setup to a single, robust gp3 volume.

Step 1: Analyze Your Current IOPS

Before resizing, check if you actually need the upgrade. Use Amazon CloudWatch to monitor the VolumeReadOps and VolumeWriteOps metrics.

# Quick check of volume stats via AWS CLI
aws cloudwatch get-metric-statistics \
    --namespace AWS/EBS \
    --metric-name VolumeWriteOps \
    --dimensions Name=VolumeId,Value=vol-0123456789abcdef0 \
    --start-time 2023-10-27T00:00:00 \
    --end-time 2023-10-28T00:00:00 \
    --period 300 \
    --statistics Sum

Step 2: Modify the Volume (Live)

One of the best features of EBS is Elastic Volumes. You can modify performance and size without detaching the volume or stopping the instance.

Let's boost our FinCore log volume to 20,000 IOPS and 1,000 MiB/s throughput to verify the performance gain.

aws ec2 modify-volume \
    --volume-id vol-0123456789abcdef0 \
    --volume-type gp3 \
    --iops 20000 \
    --throughput 1000

Step 3: Verification

Once the volume state returns to "in-use" (fully optimized), check the limits within your OS.

# Check block device details
lsblk -o NAME,SIZE,TYPE,MOUNTPOINT

# Run a quick fio benchmark (CAUTION: Do not run on prod DB without care)
fio --name=randwrite --ioengine=libaio --iodepth=1 \
  --rw=randwrite --bs=4k --direct=1 --size=1G \
  --numjobs=1 --runtime=60 --group_reporting

Conclusion

By creating a single large gp3 volume, FinCore Systems removed 3 points of failure from their storage layer and reduced their backup complexity by 75%. With the new gp3 limits, maintaining application resiliency while scaling performance is no longer a tradeoff—it's the default standard.