EBS Volume Cost Optimizations (gp2 to gp3)

Opportunity Name:

Gp2 to Gp3 migration

AWS Resource Type:

EBS Volumes

Opportunity Summary:

EBS gp2 volumes can be replaced by less expensive and equally performant gp3 volumes.

Opportunity Description:

Gp3 costs are up to 20% lower than gp2 costs, and gp3 volumes allow independent provisioning of size, IOPS, and throughput. Because gp3 is cheaper and at least as performant as gp2, gp2 volumes can almost always be converted to gp3.

Conversion is simple: a snapshot is taken of the gp2 volume, and then the volume’s type is changed to gp3.

Criteria for identifying the opportunity:

The opportunity is identified if: 

  • The EBS volume is a gp2 volume.
  • The gp2 volume is not ephemeral - that is, it has been in use for at least the last 7 days.
  • The gp2 volume is not attached to an instance which is managed by EMR.
  • Elastic Volume support is enabled for the gp2 volume.

If the volume is ephemeral then it is best dealt with by fixing the configuration which creates it.

Gp2 Volumes attached to EMR managed instances are not identified as opportunities because EMR did not support gp3 when this FF was implemented.

Potential savings (range in % on annual basis):

Up to 20% savings can be achieved by converting gp2 volumes to gp3.

What happens when the Fixer is executed?

The Fixer creates a snapshot of the volume and then changes the volume type to gp3. For large gp2 volumes that have maximum IOPS that are higher than the baseline 3000 IOPS offered by gp3, the Fixer also provisions additional IOPS for the gp3 volume to match the existing gp2 volume’s IOPS.


EBS volume modification is a safe operation, and the Fixer follows AWS best practices by creating a snapshot of the volume before modifying it. The snapshot is automatically deleted after the volume has been successfully retyped to gp3. 

Is it possible to rollback once CloudFix implements the Fixer?

In the extremely unlikely event that AWS suffers an EBS hardware failure during the retyping operation, the original volume can be restored from the snapshot manually.

Can CloudFix implement the fix automatically once I accept the recommendation?


Does this fix require downtime?


Additional Resources:

Was this article helpful?
0 out of 0 found this helpful
Have more questions? Submit a request