AWS Graviton instances in Cloudera Data Engineering

AWS Graviton is a general purpose, ARM-based processor family. AWS Graviton delivers currently the best price performance for cloud workloads running in AWS Elastic Compute Cloud (EC2). With AWS Graviton, you can optimize costs and achieve better performance. From the AWS Graviton family, Cloudera Data Engineering supports AWS Graviton 3.

Prerequisites for using AWS Graviton in Cloudera Data Engineering🔗

For information on the prerequisites, such as the Data Lake and the Spark supported versions, see Compatibility for Cloudera Data Engineering and Runtime components.

If your Data Lake version is not supported by Graviton, see Upgrading a Data Lake.

Considerations for using AWS Graviton in your cluster🔗

By default, AWS Graviton is available to all customers for Cloudera Data Engineering service compute nodes if the component version requirements are met.

To use AWS Graviton, when enabling a Cloudera Data Engineering service, select from the available Graviton instances listed under the Workload type drop-down list.

With AWS Graviton, you can use SSD instances and spot instances.

If you select the AWS Graviton workload when creating the service:

All Virtual Clusters will run AWS Graviton. Running a subset of Virtual Clusters or jobs on non-Graviton instances is not supported.
To check the list of Spark versions supported by Graviton, see Compatibility for Cloudera Data Engineering and Runtime components.