Posted On: Jan 28, 2024

Today, we are excited to launch the general availability of Interactive Endpoints for Amazon EMR on EKS. With this launch, Amazon EMR on EKS customers will be able to run interactive workloads using an integrated development environment such as EMR Studio. For customers that require control on their execution environment, they will be able to use their self-hosted Jupyter notebooks as an another mechanism to run their interactive workloads via Interactive Endpoints.

Interactive Endpoints also enhance the flexibility when running interactive workloads with Amazon EMR on EKS. When using Interactive Endpoints, customers will also be able to build a custom image with their application dependencies and use that image in running their interactive jobs using Interactive Endpoints. This is especially useful for customers who may need to use libraries in their code, which are not available in the public distribution of Amazon EMR on EKS Spark runtime. Finally, for enhanced application resiliency and accessibility, Interactive Endpoints also provides customers the full control on the instance type where the JEG pod will be deployed, including the ability to specify on-demand instances via a managed or self-managed node group. 

Amazon EMR on EKS Interactive Endpoints are generally available on Amazon EMR on EKS release version 6.10 and later, and is available in the Amazon Web Services China (Beijing) Region, operated by Sinnet, and the Amazon Web Services China (Ningxia) Region, operated by NWCD. To get started, please see our Amazon EMR on EKS Interactive Endpoints documentation.