Promhub Logo Prom hub

NVIDIA DCGM Exporter Dashboard
28,706,992 3.5 (8 reviews)

Created 5/6/2020
Updated 9/23/2021
Revision 2
Grafana Version >=6.7.3
Datasources
Prometheus

Description

This Grafana dashboard monitors NVIDIA GPU health and performance by aggregating DCGM-exported metrics into a cohesive view of temperature, power, clock speeds, memory usage, and compute utilization. It highlights both real-time hardware states and profiling data (including tensor activity) to help diagnose thermal risks, power budgets, and overall GPU throughput, with panels focused on per-GPU metrics and totalized usage.

Screenshots

Dashboard screenshot 1
Source Grafana.com

Used Metrics 7

  • DCGM_FI_DEV_FB_USED

  • DCGM_FI_DEV_GPU_TEMP

  • DCGM_FI_DEV_GPU_UTIL

  • DCGM_FI_DEV_POWER_USAGE

  • DCGM_FI_DEV_SM_CLOCK

  • DCGM_FI_PROF_PIPE_TENSOR_ACTIVE

  • gpu

Get Dashboard✕
Download
Copy to Clipboard

Dashboard Preview - NVIDIA DCGM Exporter Dashboard

Most Grafana dashboards are not optimized for small screens. For the best experience, view this preview on a larger screen.
grafana.shipit.dev/nvidia-dcgm-exporter-dashboard-1