AWS sagemaker documentation change
Summary
Updated documentation for HyperPod EKS clusters to clarify continuous provisioning as optional, describe new desired-state architecture, add link to operations guide, and fix AWS CLI command syntax
Security assessment
Changes focus on architectural descriptions and operational guidance without mentioning security vulnerabilities or access controls. The CLI command fix appears to be a typo correction rather than a security fix.
Diff
diff --git a/sagemaker/latest/dg/sagemaker-hyperpod-scaling-eks.md b/sagemaker/latest/dg/sagemaker-hyperpod-scaling-eks.md index 2830f9d8a..eccb4184d 100644 --- a//sagemaker/latest/dg/sagemaker-hyperpod-scaling-eks.md +++ b//sagemaker/latest/dg/sagemaker-hyperpod-scaling-eks.md @@ -13 +13 @@ Amazon SageMaker HyperPod clusters created with Amazon EKS orchestration now sup -Continuous provisioning is available for HyperPod clusters created with EKS orchestration. Clusters created with Slurm orchestration use a different scaling model. +Continuous provisioning is available as an optional configuration for HyperPod clusters created with EKS orchestration. Clusters created with Slurm orchestration use a different scaling model. @@ -17 +17 @@ Continuous provisioning is available for HyperPod clusters created with EKS orch -Continuous provisioning operates through an event-driven architecture that manages each instance independently. When you create a HyperPod cluster, you specify the desired number of instances for each instance group. The continuous provisioning system: +The continuous provisioning system introduces a desired-state architecture that replaces the traditional request-based model. This new architecture enables parallel, non-blocking operations across different resource levels while maintaining system stability and performance. The continuous provisioning system: @@ -74 +74 @@ Billing does not start for instances that fail to launch. If an instance launch -You must have an existing Amazon EKS cluster configured with VPC networking and the required Helm chart installed. Additionally, prepare a lifecycle configuration script and upload it to an Amazon S3 bucket that your execution role can access. +You must have an existing Amazon EKS cluster configured with VPC networking and the required Helm chart installed. Additionally, prepare a lifecycle configuration script and upload it to an Amazon S3 bucket that your execution role can access. For more information, see [Managing SageMaker HyperPod clusters orchestrated by Amazon EKS](./sagemaker-hyperpod-eks-operate.html). @@ -79 +79 @@ The following AWS CLI operation creates a HyperPod cluster with one instance gro - aws sagemaker-dev create-cluster \ + aws sagemaker create-cluster \