hyperpod-cluster-debugger

Diagnose SageMaker HyperPod cluster creation failures by running a diagnostic script that checks security groups, subnets, IAM roles, VPC endpoints, and S3 lifecycle scripts to identify configuration issues. It then provides detailed remediation steps and suggested commands to fix CloudFormation errors.

Troubleshooting EKS access and node replacement on SageMaker HyperPod clusters requires aws, jq, python3, bash, kubectl, and session-manager-plugin. These dependencies enable automated diagnostics, EFA health checks, and direct remediation of access problems.

Yes, you can validate VPC endpoints and IAM roles before creating a HyperPod cluster. The Skill performs pre-flight validation checks on security group configurations, subnet assignments, and S3 lifecycle scripts to prevent creation failures.

HyperPod cluster lifecycle scripts fail during node setup due to misconfigured S3 paths or IAM permissions. The diagnostic tool validates S3 lifecycle scripts and IAM roles, then outputs automated remediation steps to resolve the failures.

Fix EFA health check failures on Slurm orchestrated HyperPod clusters by running the diagnostic script with the cluster name and region. It identifies EFA issues and provides specific suggested commands to automate remediation.

Yes, this diagnostic tool works with both EKS and Slurm orchestration for SageMaker HyperPod clusters. It diagnoses access issues, node replacement challenges, and configuration errors across both orchestration environments.