Skip to main content

Installation and Setup

This guide covers different methods to install and set up Kubeflow in your environment.

Prerequisites​

Before installing Kubeflow, ensure you have:

  • Kubernetes cluster (version 1.25 or later)
  • kubectl configured to communicate with your cluster
  • Minimum resources: 4 CPU cores, 8 GB RAM, 50 GB disk space
  • Storage class for persistent volumes

Installation Methods​

Option 1: Kubeflow on Minikube (Local Development)​

For local development and testing:

# Start Minikube with sufficient resources
minikube start --cpus=4 --memory=8192 --disk-size=50g

# Install Kubeflow using kustomize
export KUBEFLOW_VERSION=1.8.0
export KUBEFLOW_URL="https://github.com/kubeflow/manifests/archive/v${KUBEFLOW_VERSION}.tar.gz"

# Download and extract manifests
wget -O kubeflow.tar.gz ${KUBEFLOW_URL}
tar -xvf kubeflow.tar.gz
cd manifests-${KUBEFLOW_VERSION}

# Install all Kubeflow components
while ! kustomize build example | kubectl apply -f -; do
  echo "Retrying to apply resources"
  sleep 10
done

# Wait for all pods to be ready
kubectl wait --for=condition=Ready pods --all -n kubeflow --timeout=600s

Option 2: Kubeflow on Cloud Platforms​

Google Cloud Platform (GKE):

# Create GKE cluster
gcloud container clusters create kubeflow-cluster \
  --zone=us-central1-a \
  --machine-type=n1-standard-4 \
  --num-nodes=3 \
  --enable-autoscaling \
  --min-nodes=3 \
  --max-nodes=10

# Get credentials
gcloud container clusters get-credentials kubeflow-cluster --zone=us-central1-a

# Install Kubeflow
export PIPELINE_VERSION=1.8.0
kubectl apply -k "github.com/kubeflow/pipelines/manifests/kustomize/cluster-scoped-resources?ref=$PIPELINE_VERSION"
kubectl wait --for condition=established --timeout=60s crd/applications.app.k8s.io
kubectl apply -k "github.com/kubeflow/pipelines/manifests/kustomize/env/platform-agnostic-pns?ref=$PIPELINE_VERSION"

AWS (EKS):

# Create EKS cluster
eksctl create cluster \
  --name kubeflow-cluster \
  --version 1.28 \
  --region us-west-2 \
  --nodegroup-name standard-workers \
  --node-type m5.xlarge \
  --nodes 3 \
  --nodes-min 3 \
  --nodes-max 10 \
  --managed

# Install Kubeflow
export KUBEFLOW_VERSION=1.8.0
kubectl apply -k "github.com/kubeflow/manifests/example?ref=v${KUBEFLOW_VERSION}"

Option 3: Using Kubeflow Operator​

The simplest installation method using the Kubeflow operator:

# Install the operator
kubectl apply -f https://raw.githubusercontent.com/kubeflow/kubeflow/master/bootstrap/kubeflow-operator.yaml

# Create Kubeflow instance
cat <<EOF | kubectl apply -f -
apiVersion: kubeflow.org/v1
kind: Kubeflow
metadata:
  name: kubeflow
  namespace: kubeflow
spec:
  version: v1.8.0
  components:
    - notebooks
    - pipelines
    - kfserving
    - katib
    - training
EOF

Accessing Kubeflow Dashboard​

After installation, access the Kubeflow dashboard:

# Port forward to access the dashboard
kubectl port-forward -n istio-system svc/istio-ingressgateway 8080:80

# Access at http://localhost:8080

For production, configure an Ingress or LoadBalancer service.

Next Steps​

With Kubeflow installed, you can now: