Why it matters
RDS engines manage the processor architecture for you, so moving to Graviton is usually just a change of instance class with no application changes.How nOps finds it
This recommendation comes from AWS Compute Optimizer through Cost Optimization Hub, not from an nOps rule. AWS looks at the instance’s utilization over its lookback period (14 days by default) and recommends a Graviton class that fits. The saving is AWS’s estimate, which takes your Reserved Instance and Savings Plans pricing into account.Before you start
AWS Compute Optimizer and Cost Optimization Hub enabled. The DB engine version must support Graviton classes. Permission to modify the instance (rds:ModifyDBInstance).
How to implement
- Check that your engine and version support the recommended Graviton class (RDS console: Modify → DB instance class lists only supported classes).
- Test the new class on a non-production copy if you can.
- Change the instance class in a maintenance window (RDS console → Databases → Modify → DB instance class, or
aws rds modify-db-instance --db-instance-identifier DB_ID --db-instance-class NEW_CLASS --apply-immediately). - For Aurora, change one reader at a time and then fail over, so the writer changes last.
- Update Terraform or CloudFormation so a later deployment does not change it back.
Risks and considerations
- Changing the instance class restarts the database. With Multi-AZ the outage is usually a short failover.
- Reserved Instances bought for the old class family do not cover the Graviton family.
- To roll back, modify the instance back to its original class.
FAQs
Do I need to change my application?
Do I need to change my application?
Usually not. The database engine runs on Graviton for you; clients connect the same way.