Quick Start Guide

Get KubeStellar A2A up and running in 5 minutes! This guide will walk you through the essential steps to start managing your Kubernetes clusters with AI-powered automation.

Prerequisites Check

Before we begin, make sure you have:

  • ✅ Python 3.11+ installed
  • ✅ kubectl configured with at least cluster
  • ✅ Internet connection for package downloads

Step 1: Install KubeStellar A2A

Choose your preferred installation method:

# Install uv (if not already installed)
curl -LsSf https://astral.sh/uv/install.sh | sh
 
# Clone and install
git clone https://github.com/kubestellar/a2a.git
cd a2a
uv pip install -e .

Using pip

git clone https://github.com/kubestellar/a2a.git
cd a2a
pip install -e .

Step 2: Verify Installation

Test that everything is working:

# Check installation
uv run kubestellar --help
 
# List available functions
uv run kubestellar list-functions

You should see all CLI commands:

Usage: kubestellar [OPTIONS] COMMAND [ARGS]...

Commands:
  list-functions  List all available functions
  execute        Execute a specific function
  describe       Show detailed information about a function
  agent          Start interactive AI agent

And available functions:

Available functions:

- kubestellar_management
  Description: Advanced KubeStellar multi-cluster resource management
  
- get_kubeconfig
  Description: Get details from kubeconfig file
  
- helm_deploy
  Description: Deploy Helm charts across clusters
  
- namespace_utils  
  Description: List and count resources across namespaces
  
- gvrc_discovery
  Description: Discover API resources across clusters
  
- multicluster_create
  Description: Create resources across multiple clusters
  
- multicluster_logs
  Description: Aggregate logs from multiple clusters
  
- deploy_to
  Description: Deploy resources to specific clusters

Step 3: Basic Cluster Information

Let’s start with something simple - get information about your Kubernetes clusters:

# Get basic cluster info
uv run kubestellar execute get_kubeconfig
 
# Get detailed cluster information
uv run kubestellar execute get_kubeconfig -P detail_level=full

Example output:

{
  "status": "success",
  "current_context": "kind-kubestellar",
  "total_contexts": 3,
  "clusters": [
    {
      "name": "kind-kubestellar",
      "server": "https://127.0.0.1:45243",
      "status": "accessible"
    }
  ]
}

Step 4: Explore Your Clusters

Discover what resources are available across your clusters:

# Discover all available Kubernetes resources
uv run kubestellar execute gvrc_discovery
 
# List namespaces across all clusters
uv run kubestellar execute namespace_utils -P operation=list -P all_namespaces=true

Step 5: Try Multi-Cluster Operations

Create a simple resource across multiple clusters:

# Create a ConfigMap across all accessible clusters
uv run kubestellar execute multicluster_create \
  -P resource_type=configmap \
  -P resource_name=hello-a2a \
  -P data='{"message": "Hello from KubeStellar A2A!"}' \
  -P dry_run=true

The dry_run=true flag shows what would be created without actually creating it.

Step 6: Advanced Features (Optional)

KubeStellar Management

If you have KubeStellar installed:

# Get comprehensive KubeStellar topology
uv run kubestellar execute kubestellar_management -P operation=topology_map
 
# Perform deep search with binding policy analysis
uv run kubestellar execute kubestellar_management \
  -P operation=deep_search \
  -P binding_policies=true

Helm Deployments

Deploy a simple application using Helm:

# Deploy nginx with KubeStellar binding policies
uv run kubestellar execute helm_deploy \
  -P chart_name=nginx \
  -P repository_url=https://charts.bitnami.com/bitnami \
  -P create_binding_policy=true \
  -P dry_run=true

Step 7: Try the AI Agent (Optional)

Experience natural language Kubernetes management:

# Set up OpenAI API key (if you have one)
export OPENAI_API_KEY="your-api-key-here"
 
# Start the interactive agent
uv run kubestellar agent

You’ll see the KubeStellar ASCII art:

╭─────────────────────────────────────────────────────────────────────────────────────────────╮
│  ██╗  ██╗██╗   ██╗██████╗ ███████╗███████╗████████╗███████╗██╗     ██╗      █████╗ ██████╗  │
│  ██║ ██╔╝██║   ██║██╔══██╗██╔════╝██╔════╝╚══██╔══╝██╔════╝██║     ██║     ██╔══██╗██╔══██╗ │
│  █████╔╝ ██║   ██║██████╔╝█████╗  ███████╗   ██║   █████╗  ██║     ██║     ███████║██████╔╝ │
│  ██╔═██╗ ██║   ██║██╔══██╗██╔══╝  ╚════██║   ██║   ██╔══╝  ██║     ██║     ██╔══██║██╔══██╗ │
│  ██║  ██╗╚██████╔╝██████╔╝███████╗███████║   ██║   ███████╗███████╗███████╗██║  ██║██║  ██║ │
│  ╚═╝  ╚═╝ ╚═════╝ ╚═════╝ ╚══════╝╚══════╝   ╚═╝   ╚══════╝╚══════╝╚══════╝╚═╝  ╚═╝╚═╝  ╚═╝ │
│                       Multi-Cluster Kubernetes Management Agent                             │
╰─────────────────────────────────────────────────────────────────────────────────────────────╯

Provider: openai
Model: gpt-4o

Type 'help' for available commands
Type 'exit' or Ctrl+D to quit

[openai] ▶

In the agent, try natural language commands:

# Resource queries
[openai] ▶ show me all my clusters
[openai] ▶ how many pods are running?
[openai] ▶ list all namespaces
[openai] ▶ perform deep kubestellar search

# Deployment commands
[openai] ▶ deploy nginx using helm to production
[openai] ▶ create a configmap named test-config
[openai] ▶ show me binding policies

# Troubleshooting
[openai] ▶ check cluster connectivity
[openai] ▶ find failed deployments
[openai] ▶ get logs from nginx pods

What You Just Did

🎉 Congratulations! You’ve successfully:

  • ✅ Installed KubeStellar A2A
  • ✅ Connected to your Kubernetes clusters
  • ✅ Explored available resources and namespaces
  • ✅ Performed multi-cluster operations
  • ✅ Tested advanced features like KubeStellar integration and AI automation

Next Steps

Now that you have KubeStellar A2A running, explore more advanced features:

📚 Learn More

🔧 Advanced Features

Common First Tasks

Here are some common things you might want to do next:

Deploy Your First Application

# Deploy a sample application with Helm
uv run kubestellar execute helm_deploy \
  -P chart_name=podinfo \
  -P repository_url=https://stefanprodan.github.io/podinfo \
  -P namespace=default \
  -P create_binding_policy=true
  
# Check deployment status
uv run kubestellar execute helm_deploy \
  -P operation=status \
  -P release_name=podinfo

Monitor Your Clusters

# Get logs from all pods in a namespace
uv run kubestellar execute multicluster_logs \
  -P target_namespaces='["default"]' \
  -P tail=50
 
# Check resource usage across clusters
uv run kubestellar execute namespace_utils \
  -P operation=list-resources \
  -P all_namespaces=true
  
# Stream logs in real-time
uv run kubestellar execute multicluster_logs \
  -P label_selector="app=nginx" \
  -P follow=true

Set Up Automation

# Create a script for daily cluster health checks
cat > daily-check.sh << 'EOF'
#!/bin/bash
echo "=== Daily Kubernetes Cluster Health Check ==="
echo "1. Checking cluster connectivity..."
uv run kubestellar execute get_kubeconfig -P detail_level=full
 
echo "\n2. KubeStellar topology..."
uv run kubestellar execute kubestellar_management -P operation=topology_map
 
echo "\n3. Resource inventory..."
uv run kubestellar execute gvrc_discovery
 
echo "\n4. Namespace status..."
uv run kubestellar execute namespace_utils -P operation=list -P all_namespaces=true
 
echo "\n5. Binding policy analysis..."
uv run kubestellar execute kubestellar_management -P operation=policy_analysis
 
echo "=== Health check complete ==="
EOF
 
chmod +x daily-check.sh
./daily-check.sh

CLI Parameter Examples

# Different ways to pass parameters
 
# Method 1: Using -P (recommended)
uv run kubestellar execute get_kubeconfig -P context=production -P detail_level=full
 
# Method 2: Using --param
uv run kubestellar execute get_kubeconfig --param context=production --param detail_level=full
 
# Method 3: Using JSON
uv run kubestellar execute get_kubeconfig --params '{"context": "production", "detail_level": "full"}'
 
# Complex parameters with arrays
uv run kubestellar execute helm_deploy \
  -P target_clusters='["cluster1", "cluster2"]' \
  -P set_values='["replicaCount=3", "service.type=LoadBalancer"]'
 
# Describing function parameters
uv run kubestellar describe helm_deploy
uv run kubestellar describe kubestellar_management

Troubleshooting

Installation Issues

# Verify Python version
python --version  # Should be 3.11+
 
# Check kubectl connectivity
kubectl cluster-info
 
# Verify installation
uv run kubestellar execute get_kubeconfig

Common Errors

“Function not found”: Make sure you’re using the correct function name from list-functions

“Kubeconfig not found”: Ensure kubectl is configured and $KUBECONFIG is set correctly

“Permission denied”: You might need cluster admin permissions for some operations

Getting Help

Ready to transform your Kubernetes management experience? Let’s dive deeper! 🚀