A simple reference for how we manage Intel Gaudi accelerator clusters — hardware, software, monitoring, and the path forward as the platform ages out.

• The Short Version
• Platform Status
• Setting the Record Straight on "Discontinued"
• Cluster Setup
• Keeping Software and Firmware Current
• OS Patching and Hardening
• Storage and NVMe Health
• Everyday Health Checks
• Monitoring and Diagnostics
• Security
• Bringing a New Node Online
• Keeping Configuration Consistent
• Running Shared Clusters (Multi-Tenancy)
• Handling Incidents and Making Changes
• Tracking Assets and Planning Capacity
• Backup and Disaster Recovery
• Hardware Replacement (RMA)
• Migration Planning
• New Orders
• Where to Look Things Up