AWS Batch launches new features for access control and management of AWS Batch on EKS workloads

AWS Batch on EKS now supports configurable Kubernetes namespaces, Persistent Volume Claims (PVCs), subPath to Kubernetes pod container volumes, and Kubernetes pod annotations. Using different Kubernetes namespaces for your AWS Batch jobs, you can improve workload isolation by defining job permission boundaries both within EKS cluster and for access to other AWS services. With Kubernetes PVCs and subPath you can give your AWS Batch jobs access only to the right data or particular subPath within a data volume. Finally, EKS pod annotations make it easier to integrate with external tools and other AWS services such as AWS Secrets Manager by allowing you to attach necessary metadata directly to your AWS Batch job.

You can configure Kubernetes namespaces, PVCs, subPath, and annotations while registering a new AWS Batch job definition or making a revision to an existing job definition. You can also override the namespace and annotations from your job definition when you submit the job. For more information, see Register Job Definition and Submit Job pages in the AWS Batch API reference and our AWS HPC Blog post.

AWS Batch supports developers, scientists, and engineers in running efficient batch processing for ML model training, simulations, and analysis at any scale. Configurable Kubernetes namespaces, PVCs, subPath, and annotations are available in any AWS Region where AWS Batch is available.