Move Aurora Instances to Graviton

Opportunity Name: Move Aurora Instances to Graviton

AWS Resource Type: Amazon Aurora (RDS)

Opportunity Description:
This Finder identifies provisioned Aurora database instances running on Intel-based infrastructure that are eligible to migrate to Graviton-based instances, enabling cost savings with comparable performance. Graviton processors offer improved price-performance and are fully supported for both Aurora MySQL and PostgreSQL, as well as other compatible RDS engines.

Criteria for identifying the opportunity:

• Instance is provisioned (not serverless or reserved).
• Instance is currently available and running on Intel-based infrastructure.
• Database version supports Graviton:
  • Aurora MySQL: 2.09.2 or higher
  • Aurora PostgreSQL: 12.4 or higher, 11.9 or higher
  • RDS MySQL: 8.0.17 and higher
  • RDS PostgreSQL: 13, 12.3 and higher
  • RDS MariaDB: 10.5, 10.4.13 and higher
• Instance type must support a corresponding Graviton-compatible target:
  • db.r5 or db.r6i → db.r6g
  • db.t3 → db.t4g

Potential Savings (range in % on annual basis):
On average, customers can expect ~10% savings based on current pricing differences between Intel and Graviton instance families.

What happens when the Fixer is executed?
CloudFix retypes the instance in place using the ModifyDBInstance API operation, changing the underlying instance class to a Graviton-based equivalent immediately upon approval.

Is it possible to rollback once CloudFix implements the fixer?
Yes. Rollback can be manually triggered using the SSM Document:
Cloudfix-Runbook-AuroraInstancesToGraviton-Rollback

Required parameters:

• DBClusterIdentifier
• DBInstanceIdentifier
• DBTargetClass
• IsPrimary
• ClusterNumberInstances

Can CloudFix implement the fix automatically once I accept the recommendation?
Yes. Once the recommendation is approved, CloudFix will automatically initiate the instance retyping process.

Does this fix require downtime?
Yes. Even with failover mechanisms, a short downtime (~30 seconds) is expected during the modification.
More on failover and availability from AWS Aurora FAQs

Additional Resources:

• Key Considerations for Moving to Graviton for Aurora
• AWS Aurora Documentation
• ModifyDBInstance API