Amazon SageMaker HyperPod now supports AMI versioning and auto-patching
Amazon SageMaker HyperPod provides visibility into AMI versions across clusters and automatically applies security patches without disrupting workloads, helping maintain cluster security and consistency while reducing manual patching efforts
Amazon SageMaker HyperPod is purpose-built infrastructure for training and deploying foundation models at scale. This update provides cluster administrators with visibility into the exact AMI versions running on every instance group and node in semantic versioning format (major.minor.patch), enabling quick detection of version drift and rollback to previous versions-including prior NVIDIA drivers, CUDA, and other software stacks-using the UpdateClusterSoftware API. Auto-patching is an opt-in, per-instance-group capability that applies only backward-compatible security patches as nodes become idle, ensuring running workloads remain undisturbed and critical AI/ML packages such as NVIDIA drivers, CUDA versions, and operating system kernels are never upgraded to a different major or minor version. It can be enabled through the CreateCluster or UpdateCluster API. A new AMI support policy also publishes support timelines for different AMI versions after which HyperPod stops publishing security patches. Both AMI versioning and auto-patching are available for HyperPod clusters orchestrated by Amazon EKS in all AWS Regions where SageMaker HyperPod is supported.