Introducing managed cluster extensions for observability and user management in Amazon SageMaker HyperPod Slurm
Machine Learning Blog
This article introduces managed cluster extensions for Amazon SageMaker HyperPod Slurm clusters, enabling declarative management of software and settings on running nodes through APIs and the console.
- Cluster extensions provide validated, consistent software deployment across all nodes including new and replaced ones
- Observability extension collects node, GPU, EFA, and Slurm metrics, sending to CloudWatch or Amazon Managed Service for Prometheus
- User management extension creates users and groups with consistent IDs, controls SSH and sudo access, and integrates with Active Directory
- Extensions are updated through APIs or console without replacing nodes or rerunning scripts
- Health monitoring reports per-node issues grouped by cause with codes and affected node counts
- Complements lifecycle scripts and custom AMIs for organization-specific software while managing common cluster capabilities
- Available since November 2026 in all SageMaker HyperPod regions with no additional charges
Cluster extensions bring declarative, validated management to the software layer of Slurm clusters, reducing configuration drift and operational complexity while maintaining full node access for administrators.
The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.
Related articles
2026
2026
2026
2026
The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.