AWS Launches Managed Cluster Extensions for SageMaker HyperPod Slurm Observability and User Management
AWS ML Blog · rss · 2026-11-21
AWS announced cluster extensions for Amazon SageMaker HyperPod Slurm clusters, bringing declarative API management to software running on live nodes—the sixth config layer after the five declarative layers introduced earlier in 2026.
- Problem: lifecycle scripts only run at provisioning, requiring node replacements for changes, with no status reporting, causing node drift over time
- Mechanics: schema-validated settings rejected before reaching nodes; applied consistently to all nodes including scale-ups; API/console-driven updates; per-node health reporting; full SSH access retained
- First two extensions: sagemaker-observability (node/GPU/EFA/Slurm metrics to CloudWatch or Managed Prometheus, Grafana dashboard) and sagemaker-user-management (consistent user/group IDs, SSH/sudo control, Slurm accounting, Active Directory integration)
- Six extension APIs enable scripted installs, version control, and cross-cluster repetition; custom AMIs and lifecycle scripts remain the path for org-specific software
More from Infra
- 2M canary tokens expose stolen prompts and compromised AI providers — AccBalanced · 2026-11-21
- Unsloth release runs 26.4B Inkling 2 Mini on 16GB RAM, cuts MCP token use 87% — danielhanchen · 2026-11-21
- Seroter Daily Reading List #914: harnesses are cheap to swap, model behavior isn't — rseroter · 2026-11-21
- NVFP4 MoE experts force CPU fallback in llama.cpp; requantizing boosts Metal speed 5.8x — teortaxesTex · 2026-11-21
- Inkling 2 Mini tops intelligence-per-joule test: 82.3% accuracy at just 1.17 kJ per query — simonguozirui · 2026-11-21
- Compute overbuild skeptic recants: efficiency gains cut AI compute cost ~4x per year — teortaxesTex · 2026-11-21