Amazon Redshift, a leading cloud data warehouse solution, has introduced Graviton-based RG instances, offering enhanced performance and cost efficiency. These new instances, launched on May 12, 2026, promise significant improvements in speed and pricing for data warehouse and lake workloads.
Migration Strategies: Migrating from RA3 to RG can be approached through three main strategies: Elastic Resize, Classic Resize, and Snapshot/Restore. Each method has its own considerations and best practices, ensuring a smooth transition.
Why Migrate?
Transitioning to RG instances is recommended for all RA3 users to leverage improved performance and reduced costs. RG instances maintain feature parity with RA3, allowing for seamless migration without functionality loss.
Instance Types Comparison
The RG family currently includes two node types. The following table outlines the RG instance types along with their hardware specifications and equivalent RA3 types:
| RG Instance Type | Hardware Specs | Equivalent RA3 Type |
|---|---|---|
| RG.4xlarge | 4 vCPUs, 32 GiB RAM | RA3.4xlarge |
| RG.xlarge | 2 vCPUs, 16 GiB RAM | RA3.xlarge |
Choosing Node Types: When sizing clusters, consider whether to use fewer larger nodes or more smaller nodes. Larger nodes offer greater local SSD cache, enhancing performance for I/O-intensive queries.
Migration Methods Explained
Elastic Resize: This is the preferred method when the target RG configuration is compatible. It allows for node type changes and cluster resizing while creating a snapshot of the source cluster for data transfer.
Classic Resize: Use this method when Elastic Resize is not feasible. It involves creating a new target cluster and migrating data through a backup and restore operation, which temporarily makes the source cluster unavailable.
Considerations During Migration
During the migration process, certain conditions apply:
- AWS DMS tasks are unaffected and will resume after the cluster is available.
- Zero-ETL tables may become temporarily unavailable and require resynchronization.
- Data sharing will be paused during the transition but will resume automatically post-migration.
Post-Migration Steps
After completing the migration, it is crucial to delete any test clusters and snapshots to avoid unnecessary charges. Always ensure that the correct clusters are being deleted to prevent data loss.
This guide has outlined the essential strategies and considerations for migrating Amazon Redshift RA3 instances to RG instances. For further insights on performance benefits, refer to the official announcement.
