Compute capacity for Amazon Redshift Serverless - Amazon Redshift

Compute capacity for Amazon Redshift Serverless

Understanding Amazon Redshift Serverless capacity

RPUs

Amazon Redshift Serverless measures data warehouse capacity in Redshift Processing Units (RPUs). RPUs are resources used to handle workloads.

Base capacity

This setting specifies the base data warehouse capacity Amazon Redshift uses to serve queries. Base capacity is specified in RPUs. You can set a base capacity in Redshift Processing Units (RPUs). One RPU provides 16 GB of memory. Setting higher base capacity improves query performance, especially for data processing jobs that consume a lot of resources. The default base capacity for Amazon Redshift Serverless is 128 RPUs. You can adjust the Base capacity setting from 8 RPUs to 512 RPUs in units of 8 (8,16,24...512), using the AWS console, the UpdateWorkgroup API operation, or update-workgroup operation in the AWS CLI.

With a minimum capacity of 8 RPU, you now have more flexibility to run simpler to more complex workloads based on performance requirements. The 8, 16, and 24 RPU base RPU capacities are targeted towards workloads that require less than 128TB of data. If your data requirements are greater than 128 TB, you must use a minimum of 32 RPU. For workloads that have tables with large number columns and higher concurrency, we recommend using 32 or more RPU.

Considerations and limitations for Amazon Redshift Serverless capacity

The following are considerations and limitations for Amazon Redshift Serverless capacity.

  • Configurations of 8 or 16 RPU support Redshift managed storage capacity of up to 128 TB. If you're using more than 128 TB of managed storage, you can't downgrade to less than 32 RPU.

  • Editing your workgroup's base capacity might cancel some of the queries running on your workgroup.

AI-driven scaling and optimization (preview)

This is prerelease documentation for AI-driven scaling and optimizations in Amazon Redshift Serverless, which is in preview release. The documentation and the feature are both subject to change. We recommend that you use this feature only in test environments, and not in production environments. For preview terms and conditions, see Betas and Previews in AWS Service Terms.

This preview is available in the following AWS Regions:

  • US East (Ohio) (us-east-2)

  • US East (N. Virginia) (us-east-1)

  • US West (N. California) (us-west-1)

  • Asia Pacific (Tokyo) (ap-northeast-1)

  • Europe (Ireland) (eu-west-1)

  • Europe (Stockholm) (eu-north-1)

You can create a preview workgroup to test new features of Amazon Redshift Serverless. You can't use those features in production move your workgroup to another workgroup. For preview terms and conditions, see Beta and Previews in AWS Service Terms. For instructions on how to create a preview workgroup, see Creating a preview workgroup.

You can also set a price-performance target for your workgroup, so that Redshift can automatically make AI-driven optimizations to your resources. In that way, you can meet your price-performance targets while optimizing cost. This automatic price-performance optimization is especially helpful if you don't know what base capacity to set for your workloads, or if some parts of your workload might benefit from more allocated resources.

For example, if your organization typically runs workloads that only require 32 RPU but suddenly introduces a more complex query, you might not know the appropriate amount of base capacity. Setting a higher base capacity yields better price-performance but also incurs higher costs, so the cost might not match your expectations. Using AI-driven scaling and resource optimization, Amazon Redshift Serverless automatically adjusts the RPUs to meet your price-performance targets while keeping costs optimized for your organization. This automatic optimization is useful regardless of workload size. The automatic optimization can help you meet your organization's price-performance targets if you have any number of complex queries.

Price-performance targets are a workgroup-specific setting. Different workgroups can have different price-performance targets.

To keep costs predictable, set a limit of maximum capacity that Amazon Redshift Serverless is allowed to allocate to your workloads.

To configure price-performance targets, use the AWS console. By default, price-performance target is enabled when you create a new workgroup and is set to Balanced. To set a different price-performance target or specify a base capacity for your workgroup, use customized settings when creating a workgroup. For more information about creating a workgroup, see Creating a workgroup with a namespace.

To edit the price-performance target for your workgroup:

  1. On the Amazon Redshift Serverless console, choose Workgroup configuration.

  2. Choose the workgroup for which you want to edit the price-performance target. Choose the Performance tab, then choose Edit.

  3. Choose Price-performance target, and adjust the slider according to the target you want to set your workgroup to.

  4. Choose Save changes.

To update the max amount of RPUs Amazon Redshift Serverless can allocate to your workload, go to the Limits tab of workgroup configuration.

To learn more about AI-driven optimizations and resource scaling, watch the following video.