For instructions about accessing CoreWeave Grafana dashboards, see Access and use CoreWeave Grafana dashboards.
| Panel | Description |
|---|---|
| Any Production GPUs | Total count of nodes in racks with any number of healthy nodes. |
| NVL64 Production GPUs | Baseline count of nodes based on the number of racks with at least 16 healthy nodes (Rack Count * Threshold (16)). |
| NVL64+ Production GPUs | Total count of nodes in racks with at least 16 healthy nodes. |
| NVL72 Production GPUs | Total count of nodes in racks with 18 healthy nodes. |
| NVL64+ Active Production Percentage | Percentage of active production GPUs of type NVL64 or higher. |
| Production GPUs | Distribution of the classes of Production GPUs. |
| Number of Racks with Production Schedulable Nodes | Number of racks with schedulable Nodes for production workloads. |
| Number of Racks with Production Nodes (stacked) | Stacked chart of the data in the Number of Racks with Production Nodes panel. |
GPU Utilization (max by rack_name) | Time-series graph of maximum GPU utilization aggregated by rack name. |
| Cabinet Details | Detailed table with Summary and NLCC State Counts tabs. Shows metrics such as rack name, Active status, Health rollup, and Kubernetes schedulable status. |
| Node Messages | Messages or logs related to the system’s Nodes. |