Skip to content

GPU Monitoring

Show NVIDIA or AMD GPU memory usage on the Arcane dashboard.

GPU monitoring adds a GPU Usage meter to the Dashboard for NVIDIA and AMD GPUs. It’s off by default.

  1. Install the GPU drivers on the host, following your GPU vendor’s instructions.
  2. Add the lines for your vendor (NVIDIA or AMD) to the Arcane service in compose.yaml. For a remote agent, change the agent container on the GPU host instead.
  3. Recreate the container.
  4. Check the Dashboard agent card for the GPU Usage meter.

Official Arcane images don’t bundle vendor GPU tools. NVIDIA monitoring relies on the NVIDIA container runtime injecting nvidia-smi. AMD monitoring reads /sys/class/drm directly.

services:
arcane:
environment:
- NVIDIA_VISIBLE_DEVICES=all
- ARCANE_GPU_MONITORING_ENABLED=true
- ARCANE_GPU_TYPE=nvidia
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]
services:
arcane:
environment:
- ARCANE_GPU_MONITORING_ENABLED=true
- ARCANE_GPU_TYPE=amd
devices:
- /dev/dri:/dev/dri

The meter shows average memory use across detected GPUs, with the device count below. On first collection, the logs show NVIDIA GPU detected, AMD GPU detected, or Using configured GPU type. If the meter doesn’t appear, look for these errors:

Log message Cause
nvidia-smi not found but GPU_TYPE set to nvidia The NVIDIA container runtime isn’t injecting nvidia-smi. Check the NVIDIA Container Toolkit on the host.
AMD GPU not found in sysfs but GPU_TYPE set to amd No AMD card with VRAM information was found under /sys/class/drm.
no supported GPU found Auto-detection found no GPU tool or AMD card.
Variable Default Description
ARCANE_GPU_MONITORING_ENABLED false Collect GPU stats.
ARCANE_GPU_TYPE auto nvidia, amd, intel, or auto. auto tries NVIDIA, then AMD, then Intel. Unknown values fall back to auto. Intel isn’t supported yet: Arcane detects it via intel_gpu_top but reads no memory stats, so no meter appears.