SageMaker HyperPod adds AMI versioning and auto-patching
Amazon SageMaker HyperPod now provides visibility into AMI versions and automatically applies security patches to clusters without disrupting workloads. This enhances security and consistency by enabling administrators to detect drift, roll back to previous versions, and apply patches non-disruptively. These features are available for HyperPod clusters orchestrated by Amazon EKS in all supported AWS Regions.
- →AMI versioning for SageMaker HyperPod clusters
- →Auto-patching for security updates on HyperPod
- →New AMI support policy published
- →Availability of AMI versioning and auto-patching
Features (2) ›
- AMI versioning for SageMaker HyperPod clusters
SageMaker HyperPod now offers visibility into Amazon Machine Image (AMI) versions running across clusters, presented in a semantic versioning format (major.minor.patch). This allows administrators to detect version drift and roll back to previous versions, including NVIDIA drivers and CUDA, using the UpdateClusterSoftware API.
- Auto-patching for security updates on HyperPod
An opt-in capability, auto-patching applies backward-compatible security patches to idle nodes, ensuring uninterrupted workloads. Critical AI/ML packages such as NVIDIA driver, CUDA, and OS kernels are not upgraded to different major or minor versions. This feature can be enabled via the CreateCluster or UpdateCluster API.
Enhancements (1) ›
- New AMI support policy published
A new policy defines support timelines for different AMI versions, after which SageMaker HyperPod will cease publishing security patches for them.
Notes (1) ›
- Availability of AMI versioning and auto-patching
Both AMI versioning and auto-patching are available for SageMaker HyperPod clusters orchestrated by Amazon EKS, across all AWS Regions where SageMaker HyperPod is supported.
https://aws.amazon.com/about-aws/whats-new/2026/07/amazon-sagemaker-hyperpod-ami-version-auto-patch