[WIP] Initial attempt for i2edge automation

- 2 worker nodes

- API server accessible on `0.0.0.0:6443`

- Containerd configured to keep unpacked layers

- **Port mappings** (accessible on localhost):

  - `6443` - Kubernetes API server

  - `30090` - Prometheus API (after installing dependencies)

  - `30769` - i2edge service (after deploying i2edge)

## Troubleshooting

kind create cluster --name operator-platform --config kind-config.yaml -v 3

```

## Installing i2edge Dependencies

After deploying the cluster, you can install the required dependencies for i2edge:

```bash

./install-dependencies.sh

```

This script will install:

1. **Helm3** - Package manager for Kubernetes

2. **Prometheus Stack** - Monitoring and metrics collection

   - Installed in `monitoring` namespace

   - Grafana with default credentials (admin/admin)

   - Prometheus API exposed on NodePort 30090

3. **Node Feature Discovery** - Custom node labeling

   - Installed with default configuration

   - Sample node labels for zoneID, geolocation, and geography

### Post-Installation Configuration

After running `install-dependencies.sh`, you may want to customize:

#### 1. Prometheus Configuration

Edit `prometheus-values.yaml` to add your cluster node IPs in the `endpoints` sections.

#### 2. Node Labels

Edit `nodes-labels.yaml` to customize labels for your specific nodes:

```yaml

spec:

  labels:

    feature.node.kubernetes.io/zoneID: "your-zone-id"

    feature.node.kubernetes.io/geolocation: "your-coordinates"

    feature.node.kubernetes.io/geographyDetails: "your-location-details"

```

Then reapply:

```bash

kubectl apply -f nodes-labels.yaml

```

#### 3. Access Grafana

To access Grafana dashboard:

```bash

kubectl port-forward -n monitoring svc/prometheus-grafana 3000:80

```

Then open http://localhost:3000 (login: admin/admin)

#### 4. Access Prometheus

Prometheus is exposed on NodePort 30090:

```bash

# Check the service

kubectl get svc -n monitoring prometheus-prometheus

# Access directly via localhost (port mapping configured in kind-config.yaml)

curl http://localhost:30090/api/v1/query?query=up

# Or open in browser

open http://localhost:30090

```

**Note**: With the updated `kind-config.yaml`, NodePort 30090 is directly accessible on localhost without port-forwarding!

### Verify Installation

Check that all components are running:

```bash

# Set kubeconfig

export KUBECONFIG=$(pwd)/operator-platform-external-kubeconfig.yaml

# Check Prometheus pods

kubectl get pods -n monitoring

# Check NFD pods

kubectl get pods -n node-feature-discovery

# Verify node labels

kubectl get nodes -o json | jq '.items[].metadata.labels' | grep feature.node.kubernetes.io

```

## Manual Management

If you prefer manual control:

        hostPort: 6443

        listenAddress: "0.0.0.0"

        protocol: TCP

      # Prometheus NodePort

      - containerPort: 30090

        hostPort: 30090

        listenAddress: "0.0.0.0"

        protocol: TCP

      # i2edge service NodePort

      - containerPort: 30769

        hostPort: 30769

        listenAddress: "0.0.0.0"

        protocol: TCP

  - role: worker

  - role: worker

containerdConfigPatches:

# Kind Cluster with i2edge Dependencies

This directory contains automation scripts to install i2edge dependencies on a running Kind cluster.

## Prerequisites

1. A running Kind cluster (created with `../1-kind-cluster/deploy.sh`)

2. `kubectl` configured to access the cluster

3. The cluster should have the following nodes:

   - 1 control-plane node

   - 2 worker nodes

## Verify Cluster Access

Before running the installation script, verify your cluster is accessible:

```bash

kubectl get nodes

```

Expected output:

```

NAME                              STATUS   ROLES           AGE     VERSION

operator-platform-control-plane   Ready    control-plane   5m57s   v1.33.1

operator-platform-worker          Ready    <none>          5m47s   v1.33.1

operator-platform-worker2         Ready    <none>          5m47s   v1.33.1

```

## Install i2edge Dependencies

Run the installation script:

```bash

./install-dependencies.sh

```

This script installs (Steps 1, 2, and 3 from i2edge README):

### 1. Helm3

- Package manager for Kubernetes

- Required for installing Prometheus and other charts

### 2. Prometheus Stack

- Monitoring and metrics collection system

- **Namespace**: `monitoring`

- **Components**:

  - Prometheus server (exposed on NodePort 30090)

  - Grafana dashboard

  - Alertmanager

  - Node exporters

  - Kube-state-metrics

- **Credentials**: admin/admin (configurable)

### 3. Node Feature Discovery (NFD)

- Enables custom node labeling

- **Namespace**: `node-feature-discovery` (NFD operator)

- **Custom labels namespace**: `extra-node-feature`

- **Default labels**:

  - `feature.node.kubernetes.io/zoneID`

  - `feature.node.kubernetes.io/geolocation`

  - `feature.node.kubernetes.io/geographyDetails`

## Post-Installation

After installation, the script creates configuration files in the current directory:

### 1. `prometheus-values.yaml`

Prometheus stack configuration. You may want to customize:

- Add your cluster node IPs in `endpoints` sections

- Adjust storage settings for persistence

- Modify retention policies

### 2. `nodes-labels.yaml`

Node labels configuration. Customize for your environment:

```yaml

apiVersion: nfd.k8s-sigs.io/v1alpha1

kind: NodeFeature

metadata:

  labels:

    nfd.node.kubernetes.io/node-name: operator-platform-control-plane

  name: features-test-master

  namespace: extra-node-feature

spec:

  labels:

    feature.node.kubernetes.io/zoneID: "Omega12345"

    feature.node.kubernetes.io/geolocation: "p41.388043_p2.114966"

    feature.node.kubernetes.io/geographyDetails: "university_city_env"

```

To update node labels:

```bash

kubectl apply -f nodes-labels.yaml

```

## Verify Installation

### Check all components:

```bash

# Prometheus stack

kubectl get pods -n monitoring

# Node Feature Discovery

kubectl get pods -n node-feature-discovery

# Verify node labels

kubectl get nodes -o json | jq '.items[].metadata.labels' | grep feature.node.kubernetes.io

```

### Access Grafana:

```bash

# Port forward to access Grafana locally

kubectl port-forward -n monitoring svc/prometheus-grafana 3000:80

```

Then open http://localhost:3000

- Username: `admin`

- Password: `admin`

### Access Prometheus:

Prometheus is exposed on NodePort 30090 for i2edge to query metrics.

```bash

# Check the service

kubectl get svc -n monitoring prometheus-kube-prometheus-prometheus

# Access Prometheus (for local Kind cluster)

curl http://localhost:30090/api/v1/query?query=up

```

## Troubleshooting

### Helm installation fails

Ensure you have internet connectivity and proper permissions:

```bash

curl -fsSL https://raw.githubusercontent.com/helm/helm/master/scripts/get-helm-3 | bash

```

### Prometheus pods not starting

Check pod logs:

```bash

kubectl logs -n monitoring <pod-name>

```

Common issues:

- Insufficient resources: Kind may need more memory/CPU

- Storage issues: Check PVC status with `kubectl get pvc -n monitoring`

### Node labels not appearing

Node Feature Discovery may take a few moments to process labels:

```bash

# Check NFD logs

kubectl logs -n node-feature-discovery -l app.kubernetes.io/name=node-feature-discovery

# Reapply labels

kubectl apply -f nodes-labels.yaml

```

### Cannot access Prometheus on NodePort

For Kind clusters, NodePort services are accessible on the host:

```bash

# Find the exact port

kubectl get svc -n monitoring -o wide

# Test connectivity

curl -v http://localhost:30090

```

## Next Steps

After installing dependencies:

1. **Customize configuration** (optional):

   - Edit `prometheus-values.yaml`

   - Edit `nodes-labels.yaml`

   - Reapply changes as needed

2. **Install GPU Operator** (optional, Step 4 from i2edge README):

   - Only needed if you have NVIDIA GPUs

   - See main i2edge README for instructions

3. **Deploy i2edge**:

   - Use the automated deployment script (see below)

   - Or follow manual i2edge deployment instructions

## Deploy i2edge

Once dependencies are installed, you can deploy i2edge using the automated script:

```bash

./deploy-i2edge.sh

```

This script will:

1. **Build the i2edge Docker image** from the `../../i2edge` directory

2. **Load the image into the Kind cluster** so it's available to Kubernetes

3. **Deploy all i2edge components** using Kustomize:

   - MongoDB database

   - ChartMuseum (Helm repository)

   - i2edge application with persistent volumes

   - Service accounts and RBAC configurations

4. **Wait for deployments** to become ready

5. **Display access information** for the i2edge API

### What gets deployed

- **Namespace**: `i2edge`

- **Components**:

  - **i2edge**: Main application (API + business logic)

  - **MongoDB**: Database for storing i2edge data

  - **ChartMuseum**: Local Helm repository for storing application charts

- **Persistent Volumes**:

  - Config files (kubeconfig files for cluster management)

  - Artefact files (Helm charts and application packages)

  - Deployed artefacts (tracking of deployments)

  - MongoDB data

  - ChartMuseum charts storage

### Access i2edge API

After deployment, check the output for access information. Typically:

```bash

# If using NodePort service

curl http://localhost:<nodeport>/docs

# If using ClusterIP, use port-forward

kubectl port-forward -n i2edge svc/i2edge 30769:30769

# Then access: http://localhost:30769/docs

```

The Swagger UI provides interactive API documentation.

### Verify Deployment

```bash

# Check all i2edge pods

kubectl get pods -n i2edge

# Check services

kubectl get svc -n i2edge

# Check persistent volumes

kubectl get pvc -n i2edge

# View i2edge logs

kubectl logs -n i2edge deployment/i2edge -f

```

## Undeploy i2edge

To remove i2edge from the cluster:

```bash

./undeploy-i2edge.sh

```

This script will:

1. Delete all i2edge deployments and services

2. Remove persistent volume claims and volumes

3. Delete the i2edge namespace

4. Verify cleanup

Note: This does not remove the dependencies (Prometheus, NFD, Helm) or the Kind cluster itself.

## Clean Up

To remove installed dependencies:

```bash

# Remove Prometheus

helm uninstall -n monitoring prometheus

kubectl delete namespace monitoring

# Remove Node Feature Discovery

kubectl delete -k https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=v0.16.2

kubectl delete namespace extra-node-feature

# Helm will remain installed on the host system

```

To completely remove the cluster, use the undeploy script from `../1-kind-cluster/`:

```bash

cd ../1-kind-cluster/

./undeploy.sh

```

#!/bin/bash

# Script to install i2edge dependencies on the kind cluster

# This script implements steps 1, 2, and 3 from the i2edge README

set -e  # Exit on error

echo "======================================"

echo "Installing i2edge dependencies"

echo "======================================"

# Verify cluster is accessible

if ! kubectl cluster-info &> /dev/null; then

    echo "Error: Cannot access Kubernetes cluster. Make sure the kind cluster is running."

    echo "Tip: Ensure your kubeconfig is set correctly or run 'kubectl config use-context kind-operator-platform'"

    exit 1

fi

echo "✓ Cluster is accessible"

echo "Current context: $(kubectl config current-context)"

kubectl get nodes

# =============================================

# Step 1: Install Helm3

# =============================================

echo ""

echo "======================================"

echo "Step 1: Installing Helm3"

echo "======================================"

if command -v helm &> /dev/null; then

    echo "Helm is already installed: $(helm version --short)"

else

    echo "Installing Helm3..."

    curl -fsSL -o get_helm.sh https://raw.githubusercontent.com/helm/helm/master/scripts/get-helm-3

    chmod 700 get_helm.sh

    ./get_helm.sh

    rm get_helm.sh

    echo "✓ Helm3 installed successfully"

fi

helm version

# =============================================

# Step 2: Install Prometheus Community Helm Chart

# =============================================

echo ""

echo "======================================"

echo "Step 2: Installing Prometheus"

echo "======================================"

# Add helm repository

echo "Adding prometheus-community helm repository..."

helm repo add prometheus-community https://prometheus-community.github.io/helm-charts

helm repo update

# Create monitoring namespace

echo "Creating monitoring namespace..."

kubectl create namespace monitoring --dry-run=client -o yaml | kubectl apply -f -

# Create Grafana credentials

echo "Creating Grafana credentials..."

echo -n 'admin' > ./admin-user

echo -n 'admin' > ./admin-password

kubectl create secret generic grafana-admin-credentials \

    --from-file=./admin-user \

    --from-file=admin-password \

    -n monitoring \

    --dry-run=client -o yaml | kubectl apply -f -

rm -f ./admin-user ./admin-password

echo "✓ Grafana credentials created"

# Download values.yaml if it doesn't exist

if [ ! -f "prometheus-values.yaml" ]; then

    echo "Downloading prometheus values.yaml..."

    curl -L https://raw.githubusercontent.com/techno-tim/launchpad/master/kubernetes/kube-prometheus-stack/values.yml -o prometheus-values.yaml

    # Modify the values.yaml to expose Prometheus via NodePort

    echo "Configuring Prometheus NodePort..."

    # Create a patch file for NodePort configuration

    cat > prometheus-nodeport-patch.yaml << 'EOF'

prometheus:

  service:

    type: NodePort

    nodePort: 30090

EOF

    echo "✓ Prometheus values.yaml downloaded"

    echo "⚠ Note: You may need to edit prometheus-values.yaml to add your cluster node IPs in the 'endpoints' sections"

fi

# Check if prometheus is already installed

if helm list -n monitoring | grep -q prometheus; then

    echo "Prometheus is already installed. Upgrading..."

    helm upgrade -n monitoring prometheus prometheus-community/kube-prometheus-stack -f prometheus-values.yaml

else

    echo "Installing Prometheus stack..."

    helm install -n monitoring prometheus prometheus-community/kube-prometheus-stack -f prometheus-values.yaml

fi

echo "✓ Prometheus installed successfully"

# Wait for prometheus to be ready

echo "Waiting for Prometheus pods to be ready..."

kubectl wait --for=condition=ready pod -l "release=prometheus" -n monitoring --timeout=300s || true

# =============================================

# Step 3: Install Node Feature Discovery

# =============================================

echo ""

echo "======================================"

echo "Step 3: Installing Node Feature Discovery"

echo "======================================"

echo "Installing Node Feature Discovery..."

kubectl apply -k https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=v0.16.2

# Create namespace for extra node features

echo "Creating extra-node-feature namespace..."

kubectl create namespace extra-node-feature --dry-run=client -o yaml | kubectl apply -f -

# Wait for NFD to be ready

echo "Waiting for NFD pods to be ready..."

kubectl wait --for=condition=ready pod -l app.kubernetes.io/name=node-feature-discovery -n node-feature-discovery --timeout=300s || true

echo "✓ Node Feature Discovery installed successfully"

# Create a sample NodeFeature configuration

if [ ! -f "nodes-labels.yaml" ]; then

    echo ""

    echo "Creating sample nodes-labels.yaml..."

    # Get the first node name from the cluster

    NODE_NAME=$(kubectl get nodes -o jsonpath='{.items[0].metadata.name}')

    cat > nodes-labels.yaml << EOF

apiVersion: nfd.k8s-sigs.io/v1alpha1

kind: NodeFeature

metadata:

  labels:

    nfd.node.kubernetes.io/node-name: ${NODE_NAME}

  name: features-test-master

  namespace: extra-node-feature

spec:

  # Labels to be created

  labels:

    feature.node.kubernetes.io/zoneID: "Omega12345"

    feature.node.kubernetes.io/geolocation: "p41.388043_p2.114966"

    feature.node.kubernetes.io/geographyDetails: "university_city_env"

EOF

    echo "✓ Sample nodes-labels.yaml created for node: ${NODE_NAME}"

fi

# Apply the node labels

echo "Applying node labels..."

kubectl apply -f nodes-labels.yaml

echo "✓ Node labels applied"

# Verify labels

echo ""

echo "Verifying node labels..."

kubectl get nodes -o json | jq '.items[].metadata.labels' | grep zoneID || echo "⚠ Labels may take a moment to appear"

echo ""

echo "======================================"

echo "✓ All dependencies installed successfully!"

echo "======================================"

echo ""

echo "Summary:"

echo "  - Helm3: Installed"

echo "  - Prometheus: Installed in 'monitoring' namespace"

echo "    - Prometheus exposed on NodePort 30090"

echo "  - Node Feature Discovery: Installed"

echo "    - Custom labels applied to nodes"

echo ""

echo "Next steps:"

echo "  1. Review and customize prometheus-values.yaml if needed"

echo "  2. Review and customize nodes-labels.yaml for your cluster nodes"

echo "  3. Optionally install NVIDIA GPU Operator (Step 4 in README)"

echo "  4. Deploy i2edge application"

echo ""

apiVersion: nfd.k8s-sigs.io/v1alpha1

kind: NodeFeatureRule

metadata:

  name: custom-labels-rule

spec:

  rules:

    - name: "custom-zone-labels"

      labels:

        "feature.node.kubernetes.io/zoneID": "Omega12345"

        "feature.node.kubernetes.io/geolocation": "p41.388043_p2.114966"

        "feature.node.kubernetes.io/geographyDetails": "university_city_env"