SAA-C03 Practice Question 90: GP2 to GP3 Migration
AWS Certified Solutions Architect – Associate (SAA-C03) · Test 9 · question 10 of 10
A fleet of 400 EC2 instances each has a 500 GiB gp2 root and data volume. Monitoring shows the volumes rarely exceed 2,000 IOPS or 120 MiB/s. Finance has asked for a reduction in EBS spend, and the operations team will not accept downtime or a data migration.
What should the architect do?
Show answer and explanation
Answer:
gp3 is the lowest cost SSD volume type EBS offers and is priced 20 percent lower per GiB than gp2. The performance also holds up: gp3 includes a consistent baseline of 3,000 IOPS and 125 MiB/s at no extra cost regardless of volume size, which is above this fleet's observed 2,000 IOPS and 120 MiB/s, so the defaults are enough and no additional IOPS or throughput need to be purchased. AWS documents Elastic Volumes operations as the way to migrate gp2 to gp3, and they modify the volume type without interrupting the running instance, meeting the no-downtime constraint.
Option B changes performance and is not even possible in place. An EBS volume can be grown but never shrunk, so this would require creating smaller volumes and copying data, a migration with downtime. On gp2, halving the size would also halve the baseline IOPS to 750, well under what the fleet uses.
Option C swaps SSD for HDD. st1 is cheaper per GiB but is built for large sequential streaming workloads; it cannot be a root volume, and its IOPS profile would badly hurt the random access these general-purpose volumes serve.
Option D reaches the right volume type by the most disruptive route. Snapshot, delete and restore takes the instances down and risks data loss, when a live volume modification achieves the same result with no outage.
Reference: Amazon EBS General Purpose SSD volumes