Amazon SageMaker HyperPod now supports custom Kubernetes labels and taints

Amazon SageMaker HyperPod Update

Amazon SageMaker HyperPod now supports custom Kubernetes labels and taints, enabling precise control over pod scheduling and integration with existing Kubernetes infrastructure. This update allows for managed configuration of labels and taints at the instance group level, reducing operational overhead.

What to do

• Use the new KubernetesConfig parameter in CreateCluster and UpdateCluster APIs to specify up to 50 labels and 50 taints per instance group.
• Configure labels for resource organization and pod targeting, and taints to repel pods without matching tolerations.
• Apply NoSchedule taints to GPU instance groups to ensure only AI training jobs with explicit tolerations consume high-cost compute resources.

This feature is available in all AWS Regions where Amazon SageMaker HyperPod is available. To learn more, see the user guide.

If you need further guidance on AWS, our experts are available at AWS@westloop.io. You may also reach us by submitting the Contact Us form.