intel-gpu-exporter

intel-gpu-exporter

Exports GPU statistics for Intel Quick Sync Video GPUs.

Installation

Container images are available on ghcr.io.

This is how I run the container inside my Kubernetes cluster:

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: intel-gpu-exporter
spec:
  selector:
    matchLabels:
      app: intel-gpu-exporter
  template:
    metadata:
      labels:
        app: intel-gpu-exporter
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/port: "9100"
    spec:
      nodeSelector:
        intel.feature.node.kubernetes.io/gpu: "true"
      hostPID: true
      volumes:
        - name: dev-dri
          hostPath:
            path: /dev/dri
            type: Directory
      containers:
      - name: intel-gpu-exporter
        image: ghcr.io/clambin/intel-gpu-exporter
        args:
        - --interval=5s
        - --device=drm:/dev/dri/card0
        ports:
        - name: metrics
          containerPort: 9100
        securityContext:
          privileged: true
        volumeMounts:
          - name: dev-dri
            mountPath: /dev/dri

The combination of running a DaemonSet and setting the nodeSelector to intel.feature.node.kubernetes.io/gpu="true" runs the exporter on every node with an Intel Quick Sync Video GPU. This does require intel-device-plugins-for-kubernetes. You may be better served by running a Deployment with a node selector for your host(s) that have an Intel Quick Sync Video GPU.

Limitations

• Based on the intel_gpu_top project. Practically speaking, this means that it only supports Intel Quick Sync Video GPUs that are supported by the i915 driver.
• Developed using intel_gpu_top v2.3. The built container image packages this version.

Metrics

Dashboard

See dashboards.yaml for a Grafana dashboard visualizing these metrics. This dashboard is exported as a grafana-operator dashboard.

Authors

• Christophe Lambin

License

This project is licensed under the MIT license.