> ## Documentation Index
> Fetch the complete documentation index at: https://docs.visent.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Metrics

> Complete reference of GPU and system metrics collected by Visent Telemetry

# Metrics

Visent Telemetry collects comprehensive metrics about GPU performance, system resources, and application usage. All metrics include timestamps, node identifiers, and GPU device information for detailed analysis and alerting.

Metrics are collected at configurable intervals and can be filtered, aggregated, and exported to external monitoring systems.

## GPU Metrics

Coming soon - detailed list of GPU-specific metrics including utilization, memory, and performance counters.

### Core GPU Metrics

| Metric            | Description             | Unit       |
| ----------------- | ----------------------- | ---------- |
| `gpu_utilization` | GPU compute utilization | Percentage |
| `gpu_memory_used` | GPU memory usage        | Bytes      |
| `gpu_temperature` | GPU temperature         | Celsius    |
| `gpu_power_draw`  | GPU power consumption   | Watts      |

### Memory Metrics

Coming soon - GPU memory usage, allocation, and fragmentation metrics.

### Performance Metrics

Coming soon - throughput, latency, and efficiency measurements.

## System Metrics

Coming soon - host system metrics including CPU, memory, network, and storage.

## Process Metrics

Coming soon - per-process GPU usage and resource consumption.

## Custom Metrics

Coming soon - how to define and collect application-specific metrics.

## Metric Retention

Coming soon - data retention policies and storage optimization.

## Next Steps

* [Set up alerts](alerts) based on these metrics
* [Configure integrations](../integrations/grafana) for visualization
* [Use the API](../api/telemetry) to query metrics programmatically