CASE 113 · PACER · 2023
gp2 to gp3, with the IOPS provisioned for what the workload actually needs.
An engineering analytics company had 240TB of gp2 EBS volumes — most of them oversized because gp2 couples IOPS to capacity. We migrated to gp3 with IOPS sized to observed peak, and dropped the EBS bill 38%.
Engineering analytics
COST
2023
RESULTS
What changed, by the numbers.
EBS BILL
−38%
IOPS HEADROOM
+22%
CAPACITY RIGHT-SIZED
−18%
MIGRATION DOWNTIME
0
HOW IT WENT
On gp2 you provisioned capacity to get IOPS — 1TB gave you 3,000 IOPS, so workloads that needed IOPS got fat. Most teams stopped questioning the practice. The post-gp3 reality is that you size capacity for capacity and provision IOPS independently.
We mapped every gp2 volume’s observed IOPS over 90 days and modelled the right gp3 configuration. Capacity dropped on volumes that had been oversized for IOPS reasons. IOPS provisioning was right-sized to peak + 20% headroom. Elastic Volumes handled the migration in-place, no downtime.
EBS bill dropped 38%. IOPS headroom went up 22% because the new sizing was deliberate. Cumulative capacity dropped 18% because volumes that had been padded for IOPS purposes could shrink. The migration ran across 480 volumes over four weeks with zero downtime.
RELATED · SAME DOMAIN
Other engagements in this space.
READY WHEN YOU ARE
Let's get your AWS bill (and architecture) in order.
The discovery call is free. You walk away with at least one concrete idea — even if we never work together.