NVIDIA/gpu-operator

NVIDIA GPU Operator creates, configures, and manages GPUs in Kubernetes

What it solves

It eliminates the manual, error-prone process of configuring and managing NVIDIA GPU software on Kubernetes nodes. Instead of requiring a specialized OS image for every GPU node, it allows administrators to use standard OS images and automate the deployment of necessary drivers and runtimes.

How it works

Using the Kubernetes operator framework, the project automates the lifecycle of all required NVIDIA software components. It deploys these components as containers, including NVIDIA drivers (for CUDA), the Kubernetes device plugin for GPUs, the NVIDIA Container Runtime, and DCGM-based monitoring, while also handling automatic node labelling.

Who it’s for

Kubernetes cluster administrators who need to provision and scale GPU nodes quickly across cloud or on-premises environments without manually managing the underlying software stack.

Highlights

  • Automates the installation of NVIDIA drivers, container runtimes, and device plugins.
  • Enables the use of standard OS images for both CPU and GPU nodes.
  • Simplifies component updates and swaps by running software as containers.
  • Provides integrated monitoring via DCGM.

Related

  • Project
    NVIDIA/k8s-device-plugin

    The official NVIDIA device plugin for Kubernetes that enables the cluster to automatically detect, monitor, and allocate NVIDIA GPUs to containers.

  • Project
    NVIDIA/NVSentinel

    A Kubernetes-native tool for automatically detecting, isolating, and remediating hardware and software faults in NVIDIA GPU nodes to ensure infrastructure reliability.

  • Project
    kaito-project/kaito

    A Kubernetes operator suite that automates the deployment, scaling, and management of LLM inference, fine-tuning, and RAG engines.

  • Project
    ROCm/k8s-device-plugin

    A Kubernetes device plugin that enables the registration and scheduling of AMD GPUs as compute resources within a container cluster.

  • Project
    ray-project/kuberay

    A Kubernetes operator that simplifies the deployment, scaling, and management of Ray applications for distributed AI workloads.