splunk-ai-operator

Installation Guide

This guide covers all installation methods for the Splunk AI Operator.

Prerequisites

Kubernetes v1.31+ cluster
kubectl v1.11.3+
Helm v3.8+ (for Helm installation)

Installation Methods

Method 1: Helm OCI Registry (Recommended)

The recommended way to install using Helm 3.8+:

# Install operator
helm install splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 0.1.0 \
  --namespace splunk-ai-operator-system \
  --create-namespace

With custom values:

# Create values file
cat > my-values.yaml <<EOF
replicaCount: 2
image:
  pullPolicy: IfNotPresent
resources:
  limits:
    cpu: 1000m
    memory: 512Mi
EOF

# Install with custom values
helm install splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 1.0.0 \
  --namespace splunk-ai-operator-system \
  --create-namespace \
  --values my-values.yaml

View available versions:

# Visit GHCR package page
# https://github.com/splunk/splunk-ai-operator/pkgs/container/charts%2Fsplunk-ai-operator

Method 2: kubectl (Manifests)

Install using Kubernetes manifests:

# Install operator with all dependencies
kubectl apply -f https://github.com/splunk/splunk-ai-operator/releases/download/v1.0.0/install-v1.0.0.yaml

Verify installation:

kubectl get pods -n splunk-ai-operator-system
kubectl get crds | grep ai.splunk.com

Method 3: Helm from GitHub Release

For compatibility with older Helm versions (< 3.8):

# Install from GitHub Release
helm install splunk-ai-operator \
  https://github.com/splunk/splunk-ai-operator/releases/download/v1.0.0/splunk-ai-operator-1.0.0.tgz \
  --namespace splunk-ai-operator-system \
  --create-namespace

Method 4: From Source (Development)

For developers working on the operator:

# Clone repository
git clone https://github.com/splunk/splunk-ai-operator.git
cd splunk-ai-operator

# Install CRDs
make install

# Build and push image
make docker-build docker-push IMG=ghcr.io/YOUR_ORG/splunk-ai-operator:dev

# Deploy
make deploy IMG=ghcr.io/YOUR_ORG/splunk-ai-operator:dev

See Local Development Guide for more details.

Installation Scope

Cluster-Scoped (Default)

By default, the operator is installed cluster-scoped and watches all namespaces:

# Helm installation is cluster-scoped by default
helm install splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 1.0.0 \
  --namespace splunk-ai-operator-system \
  --create-namespace

Namespace-Scoped

To restrict the operator to a single namespace:

# Install with namespace scope
helm install splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 1.0.0 \
  --namespace my-app-namespace \
  --create-namespace \
  --set watchNamespace=my-app-namespace

Private Registry Support

Using Private Container Registry

If using a private registry for the operator image:

# Create image pull secret
kubectl create secret docker-registry private-registry \
  --docker-server=your-registry.com \
  --docker-username=your-username \
  --docker-password=your-password \
  --namespace splunk-ai-operator-system

# Install with private registry
helm install splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 1.0.0 \
  --namespace splunk-ai-operator-system \
  --create-namespace \
  --set image.repository=your-registry.com/splunk-ai-operator \
  --set image.tag=1.0.0 \
  --set imagePullSecrets[0].name=private-registry

Configure private registry for related images (Ray, Weaviate, etc.):

helm install splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 1.0.0 \
  --namespace splunk-ai-operator-system \
  --create-namespace \
  --set env.RELATED_IMAGE_RAY_HEAD=your-registry.com/ray-head:latest \
  --set env.RELATED_IMAGE_WEAVIATE=your-registry.com/weaviate:latest

Advanced Configuration

Custom Cluster Domain

If your cluster uses a custom domain (not cluster.local):

helm install splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 1.0.0 \
  --namespace splunk-ai-operator-system \
  --create-namespace \
  --set env.CLUSTER_DOMAIN=internal.mycluster

Resource Limits

Configure resource limits:

helm install splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 1.0.0 \
  --namespace splunk-ai-operator-system \
  --create-namespace \
  --set resources.limits.cpu=1000m \
  --set resources.limits.memory=512Mi \
  --set resources.requests.cpu=100m \
  --set resources.requests.memory=256Mi

High Availability

Run multiple replicas for high availability:

helm install splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 1.0.0 \
  --namespace splunk-ai-operator-system \
  --create-namespace \
  --set replicaCount=3 \
  --set leaderElection.enabled=true

Verification

Check Operator Status

# Check operator pods
kubectl get pods -n splunk-ai-operator-system

# Check operator logs
kubectl logs -n splunk-ai-operator-system \
  -l control-plane=controller-manager \
  --tail=100

# Verify CRDs are installed
kubectl get crds | grep ai.splunk.com

Expected output:

aiplatforms.ai.splunk.com
aiservices.ai.splunk.com

Check Webhooks

# Verify webhook configuration
kubectl get validatingwebhookconfigurations | grep splunk-ai-operator
kubectl get mutatingwebhookconfigurations | grep splunk-ai-operator

# Check certificate
kubectl get certificates -n splunk-ai-operator-system

Upgrading

Upgrade Operator

# Upgrade to new version
helm upgrade splunk-ai-operator \
  oci://ghcr.io/splunk/charts/splunk-ai-operator \
  --version 1.1.0 \
  --namespace splunk-ai-operator-system

View Upgrade Status

# Check helm release history
helm history splunk-ai-operator -n splunk-ai-operator-system

# Check rollout status
kubectl rollout status deployment/splunk-ai-operator-controller-manager \
  -n splunk-ai-operator-system

Uninstallation

Using Helm

# Uninstall operator
helm uninstall splunk-ai-operator \
  --namespace splunk-ai-operator-system

# Delete namespace (if desired)
kubectl delete namespace splunk-ai-operator-system

# Remove CRDs (optional - this will delete all custom resources)
kubectl delete crd aiplatforms.ai.splunk.com
kubectl delete crd aiservices.ai.splunk.com

Using kubectl

# Delete manifests
kubectl delete -f https://github.com/splunk/splunk-ai-operator/releases/download/v1.0.0/install-v1.0.0.yaml

# Delete namespace
kubectl delete namespace splunk-ai-operator-system

Troubleshooting

Operator Pod Not Starting

# Check pod status
kubectl describe pod -n splunk-ai-operator-system \
  -l control-plane=controller-manager

# Check events
kubectl get events -n splunk-ai-operator-system --sort-by='.lastTimestamp'

Webhook Issues

If webhook is not working:

# Check webhook service
kubectl get svc -n splunk-ai-operator-system

# Check certificate
kubectl get secret -n splunk-ai-operator-system | grep tls

# Delete and recreate webhook (cert-manager will regenerate)
kubectl delete validatingwebhookconfigurations splunk-ai-operator-validating-webhook
kubectl delete mutatingwebhookconfigurations splunk-ai-operator-mutating-webhook

Image Pull Errors

# Check image pull secrets
kubectl get secrets -n splunk-ai-operator-system

# Verify image exists
docker pull ghcr.io/splunk/splunk-ai-operator:v1.0.0

# Check pod events for pull errors
kubectl describe pod -n splunk-ai-operator-system <pod-name>

Next Steps

After installing the operator:

Deploy AI Platform: See Helm Deployment Guide
Configure Storage: See Storage Configuration
Set Up Ingress: See Ingress Configuration
Review API: See API Reference

Additional Resources

This site is open source. Improve this page.