
Key Metrics
| Metric | Description |
|---|---|
| Total GPUs | Total GPU count across all clusters |
| Running GPUs | GPUs currently executing workloads |
| Unused GPUs | GPUs not currently in use |
| Active Workloads | Running and pending workloads |
Features
- Cluster filtering: View metrics for specific clusters or all clusters
- Unused GPUs by Chip Type: Breakdown of idle GPUs by GPU model
- Active Workloads chart: Running vs queued workloads over the last 48 hours
- Utilization trends: Track GPU usage over time
- AI Insights: Get optimization recommendations

