Installation

This guide covers installing KLearn on a Kubernetes cluster for development or production use.

Requirements

Component	Minimum	Recommended
Kubernetes	1.26+	1.29+
CPU	4 cores	8+ cores
Memory	16 GB	32 GB
Storage	50 GB	100 GB

Required Components

kubectl - Kubernetes CLI
Helm 3 - Kubernetes package manager
Docker - Container runtime (for building images)

Installation Methods

Method 1: Helm Chart (Recommended)

The easiest way to install KLearn is using our Helm chart:

# Add KLearn Helm repository
helm repo add klearn https://klearn-dev.github.io/charts
helm repo update

# Create namespace
kubectl create namespace klearn

# Install KLearn
helm install klearn klearn/klearn \
  --namespace klearn \
  --values values.yaml

Method 2: Local Development with kind

For local development and testing:

# Clone the repository
git clone https://github.com/klearn-dev/klearn.git
cd klearn

# Create kind cluster with registry
./scripts/create-kind-cluster.sh

# Build and deploy
make build push deploy

Method 3: Kustomize

For GitOps workflows:

# Clone the repository
git clone https://github.com/klearn-dev/klearn.git
cd klearn

# Apply with kustomize
kubectl apply -k deploy/kubernetes/overlays/prod

Configuration

Helm Values

Create a values.yaml file to customize your installation:

# Backend configuration
backend:
  replicas: 2
  resources:
    requests:
      cpu: 500m
      memory: 512Mi
    limits:
      cpu: 2000m
      memory: 2Gi

# Frontend configuration
frontend:
  replicas: 2

# Database
postgresql:
  enabled: true
  auth:
    username: klearn
    password: your-secure-password
    database: klearn

# Object Storage
minio:
  enabled: true
  rootUser: minioadmin
  rootPassword: your-secure-password

# ML Tracking
mlflow:
  enabled: true

# LLM Service
ollama:
  enabled: true
  model: llama3.2

Environment Variables

Variable	Description	Default
`DATABASE_URL`	PostgreSQL connection string	`postgresql+asyncpg://...`
`REDIS_URL`	Redis connection string	`redis://localhost:6379/0`
`MINIO_ENDPOINT`	MinIO server address	`localhost:9000`
`MLFLOW_TRACKING_URI`	MLflow tracking server	`http://localhost:5000`
`K8S_NAMESPACE`	Kubernetes namespace	`klearn`

Verify Installation

Check that all pods are running:

# Check pods
kubectl get pods -n klearn

# Expected output:
# NAME                               READY   STATUS    RESTARTS   AGE
# klearn-backend-xxx                 1/1     Running   0          2m
# klearn-frontend-xxx                1/1     Running   0          2m
# klearn-operator-xxx                1/1     Running   0          2m
# klearn-minio-xxx                   1/1     Running   0          2m
# klearn-mlflow-xxx                  1/1     Running   0          2m
# klearn-postgresql-xxx              1/1     Running   0          2m
# klearn-redis-xxx                   1/1     Running   0          2m

Check CRDs:

kubectl get crd | grep klearn

# Expected output:
# klearnjobs.klearn.klearn.dev
# klearnmodels.klearn.klearn.dev

Troubleshooting

Pods not starting

Check pod logs:

kubectl logs -n klearn deployment/klearn-backend
kubectl logs -n klearn deployment/klearn-operator

Database connection issues

Verify PostgreSQL is running:

kubectl exec -it -n klearn deployment/klearn-postgresql -- psql -U klearn -c "SELECT 1"

Storage issues

Check MinIO connectivity:

kubectl port-forward -n klearn svc/klearn-minio 9000:9000 &
curl http://localhost:9000/minio/health/live

Next Steps

Quick Start - Train your first model
Configuration - Advanced configuration options
Architecture - Understand how KLearn works