Cloud AI Copilots

Intelligent multi-agent platform for Kubernetes management and cloud cost optimization

Built with Google Agent Development Kit (ADK) and the A2A (Agent-to-Agent) protocol for seamless inter-agent communication.

🎯 Overview

Cloud AI Copilots is a multi-agent platform that combines specialized AI agents for comprehensive cloud infrastructure management. The platform consists of three integrated services:

K8s Copilot - Kubernetes troubleshooting and management
Cost Copilot - GCP cost analysis and optimization
Unified Dashboard - Central management UI

✨ Features

K8s Copilot

🔍 Multi-agent diagnostics - Automated cluster health checks and resource monitoring
🕵️ Intelligent investigation - Root cause analysis for pod failures and deployment issues
💡 Remediation suggestions - Actionable fix recommendations with kubectl commands
💰 Integrated cost analysis - Namespace and cluster cost estimation via A2A

Cost Copilot

💸 Idle resource detection - Find unused VMs, disks, IPs, and load balancers
📊 Cost optimization - GCP billing analysis and recommendations
🎯 K8s cost estimation - Calculate namespace and cluster costs from resource requests
📈 Cost comparison - Compare costs across multiple namespaces

Unified Dashboard

🎨 Modern UI - Built with TanStack Start, React, and Tailwind CSS
⚡ Instant loading - Static configuration for fast page loads
🔗 Agent discovery - View all available agents and their capabilities
📱 Responsive design - Works on desktop and mobile devices

🏗️ Architecture

┌─────────────────────────────────────────────────────────────┐
│                    Cloud AI Copilots                        │
│                  (Microservices Architecture)                │
└─────────────────────────────────────────────────────────────┘
                          │
        ┌─────────────────┼─────────────────┬────────────────┐
        │                 │                 │                │
   ┌────▼────┐      ┌─────▼──────┐   ┌────▼────────┐  ┌────▼────┐
   │ K8s     │ A2A  │ Cost       │   │ Unified     │  │  User   │
   │ Copilot │◄────►│ Copilot    │   │ Dashboard   │  │         │
   │ :8000   │      │ :8001      │   │ :3000       │  │         │
   └─────────┘      └────────────┘   └─────────────┘  └─────────┘
        │                 │                 │
   ┌────▼────┐      ┌─────▼──────┐
   │   GKE   │      │ GCP Billing│
   │ Cluster │      │ & Assets   │
   └─────────┘      └────────────┘

A2A Integration

The K8s Copilot and Cost Copilot communicate via the A2A (Agent-to-Agent) protocol:

User: "What's the cost of my namespace?"
  │
  ├─► K8s Copilot (Diagnose namespace resources)
  │     ├─► DiagnosticAgent.get_namespace_resources()
  │     └─► Returns: 2.0 cores, 8.0 GB RAM, 4 pods
  │
  └─► Cost Copilot (via A2A)
        ├─► CostAnalysisAgent.estimate_namespace_cost()
        └─► Returns: $69.92/month breakdown

🚀 Quick Start

Prerequisites

Python 3.10+
Node.js 18+
Google Cloud SDK
Kubernetes cluster access (for K8s Copilot)
GCP project with billing enabled (for Cost Copilot)
Gemini API key

1. Clone the Repository

git clone https://github.com/m3rryqold/cloud-copilot.git
cd cloud-copilot

2. Set Up K8s Copilot

cd services/k8s-copilot
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

# Create .env file
echo "GOOGLE_API_KEY=your-gemini-api-key" > .env
echo "COST_COPILOT_URL=http://localhost:8001" >> .env

# Start the agent
adk web src/agents --port 8000 --host 0.0.0.0

Visit: http://localhost:8000

3. Set Up Cost Copilot

cd services/cost-copilot
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

# Create .env file
echo "GOOGLE_API_KEY=your-gemini-api-key" > .env
echo "GCP_PROJECT_ID=your-gcp-project-id" >> .env

# Start the A2A server
python start_a2a.py

A2A endpoint: http://localhost:8001

4. Set Up Unified Dashboard

cd services/unified-dashboard
npm install

# Create .env.local file
cat > .env.local <<ENVEOF
VITE_K8S_COPILOT_URL=http://localhost:8000
VITE_COST_COPILOT_URL=http://localhost:8001
ENVEOF

# Start the dashboard
npm run dev

Visit: http://localhost:3000

📖 Usage Examples

K8s Copilot

💬 "Check the health of my cluster"
💬 "Why is my pod in CrashLoopBackOff?"
💬 "Show me the cost of the production namespace"
💬 "Compare costs between staging and production namespaces"

Cost Copilot

💬 "Find all idle VMs in my project"
💬 "Show me unattached disks that are costing money"
💬 "Estimate the cost of my Kubernetes cluster"
💬 "What are my top 10 most expensive services?"

Unified Dashboard

Browse available agents and their capabilities
View agent architecture and communication flows
Quick access to agent web UIs
Responsive design for mobile and desktop

🛠️ Technology Stack

Backend (Python)

Google ADK - Agent framework
Gemini 2.5 - Large language model
Kubernetes Python Client - K8s API interaction
Google Cloud Client Libraries - GCP services
Uvicorn - ASGI server

Frontend (TypeScript)

TanStack Start - Full-stack React framework
React 18 - UI library
TypeScript - Type safety
Tailwind CSS - Styling
shadcn/ui - Component library
Vite - Build tool

📂 Project Structure

cloudcopilot/
├── services/
│   ├── k8s-copilot/          # Kubernetes management agent
│   │   ├── src/agents/
│   │   ├── requirements.txt
│   │   └── .env
│   ├── cost-copilot/         # Cost optimization agent
│   │   ├── src/agents/
│   │   ├── start_a2a.py
│   │   ├── requirements.txt
│   │   └── .env
│   └── unified-dashboard/    # Web dashboard
│       ├── src/
│       ├── public/
│       ├── package.json
│       └── .env.local
├── docs/                     # Documentation
├── README.md                 # This file
├── CLAUDE.md                 # Development context
└── .gitignore

🔧 Configuration

K8s Copilot (.env)

GOOGLE_API_KEY=your-gemini-api-key
COST_COPILOT_URL=http://localhost:8001  # or Cloud Run URL

Cost Copilot (.env)

GOOGLE_API_KEY=your-gemini-api-key
GCP_PROJECT_ID=your-gcp-project-id
PUBLIC_HOST=localhost
PUBLIC_PORT=8001
PROTOCOL=http

Unified Dashboard (.env.local)

VITE_K8S_COPILOT_URL=http://localhost:8000
VITE_COST_COPILOT_URL=http://localhost:8001

🚢 Deployment

Google Cloud Run

# Deploy all services via Cloud Build
gcloud builds submit --config cloudbuild.yaml --project=your-project-id

# Services will be deployed to:
# - https://k8s-copilot-xxx.run.app
# - https://cost-copilot-xxx.run.app
# - https://unified-dashboard-xxx.run.app

🧪 Testing

Verify A2A Connection

# Check Cost Copilot agent card
curl http://localhost:8001/.well-known/agent-card.json | python3 -m json.tool

# Should return agent metadata with skills list

Test Namespace Cost Query

1. Start both K8s Copilot and Cost Copilot
2. In K8s Copilot UI: "Tell me the cost of my default namespace"
3. Expected: Cost breakdown with cores, memory, and estimated monthly cost

📝 Development

See CLAUDE.md for detailed development context and guidelines.

Adding New Tools

K8s Copilot:

# In services/k8s-copilot/src/agents/diagnostic_agent/tools.py
@tool
def your_new_tool(param: str) -> str:
    """Tool description for the LLM."""
    return result

Cost Copilot:

# In services/cost-copilot/src/agents/cost_analysis_agent/tools.py
@tool
def your_new_cost_tool(project_id: str) -> dict:
    """Tool description for the LLM."""
    return result

Dashboard:

// In services/unified-dashboard/src/config/agents.ts
export const AGENT_CONFIGS: AgentConfig[] = [
  // Add your new agent configuration
]

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

This project is licensed under the MIT License.

🙏 Acknowledgments

Built with Google Agent Development Kit (ADK)
Powered by Gemini 2.5
UI components from shadcn/ui

📬 Support

For issues and questions:

Create an issue in this repository
Review the A2A Integration docs

Built with ❤️ using Google ADK and A2A Protocol

Name		Name	Last commit message	Last commit date
Latest commit History 18 Commits
docs		docs
scripts		scripts
services		services
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
cloudbuild.yaml		cloudbuild.yaml

Folders and files

Latest commit

History

Repository files navigation

Cloud AI Copilots

🎯 Overview

✨ Features

K8s Copilot

Cost Copilot

Unified Dashboard

🏗️ Architecture

A2A Integration

🚀 Quick Start

Prerequisites

1. Clone the Repository

2. Set Up K8s Copilot

3. Set Up Cost Copilot

4. Set Up Unified Dashboard

📖 Usage Examples

K8s Copilot

Cost Copilot

Unified Dashboard

🛠️ Technology Stack

Backend (Python)

Frontend (TypeScript)

📂 Project Structure

🔧 Configuration

K8s Copilot (.env)

Cost Copilot (.env)

Unified Dashboard (.env.local)

🚢 Deployment

Google Cloud Run

🧪 Testing

Verify A2A Connection

Test Namespace Cost Query

📝 Development

Adding New Tools

🤝 Contributing

📄 License

🙏 Acknowledgments

📬 Support

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages