Skip to content

QUICK_SETUP.md

Latest commit

 

History

History
260 lines (207 loc) · 6.83 KB

QUICK_SETUP.md

File metadata and controls

260 lines (207 loc) · 6.83 KB
Raw

⚡ QUICK SETUP - Modern AWS DevOps Agent Ready in 3 Minutes

1️⃣ Install Dependencies

# Install UV (if not installed)
curl -LsSf https://astral.sh/uv/install.sh | sh

# Install Python dependencies
pip install -r requirements.txt

# For development (optional)
pip install -r requirements_dev.txt

2️⃣ Setup MCP Servers (Choose One)

Option A: Install for Production (Recommended)

# Install MCP servers globally (AWS + GitHub)
make mcp-install

# Check status
make mcp-check

GitHub MCP Configuration (Optional):

# Add GitHub token for repository management
echo "GITHUB_PERSONAL_ACCESS_TOKEN=ghp_your_token_here" >> src/aws_devops_agent/config/.env

# Test GitHub connectivity
make github-test-connectivity REPO=octocat/Hello-World

Option B: Run Directly for Development

# Run MCP servers directly (no installation, always latest)
make mcp-run

# Stop when done
make mcp-stop

3️⃣ Configure Environment

For AgentCore Deployment (Production):

# Create production environment
make agentcore-env-prod

# Validate configuration
make agentcore-validate

For Development:

# Create development environment
make agentcore-env-dev

# Or copy environment template
cp .env.example .env

# Configure AWS credentials (choose one method):

# Method 1: AWS CLI
aws configure

# Method 2: Environment variables
export AWS_ACCESS_KEY_ID=your_key
export AWS_SECRET_ACCESS_KEY=your_secret  
export AWS_DEFAULT_REGION=us-east-1

# Method 3: Edit .env file
# Edit .env file with your credentials

3.5️⃣ CDK Setup (Optional - for CDK Analysis)

# Install AWS CDK CLI (if not already installed)
npm install -g aws-cdk

# Verify CDK installation
cdk --version

# Bootstrap CDK (if using CDK for the first time in your AWS account)
cdk bootstrap

4️⃣ Test MCP Integration

python tests/test_mcp_integration.py

5️⃣ Run the Agent

🚀 EASY WAY (Recommended)

# Complete setup and run (BEST)
make setup
source .venv/bin/activate
make run

# Or step by step
make setup            # Create venv, install deps, install MCP servers
source .venv/bin/activate  # Activate venv
make run              # Start the agent

🔧 Manual Way

# Development mode (requires virtual environment)
source .venv/bin/activate
PYTHONPATH=src python main.py --mode interactive

# Demo mode
source .venv/bin/activate
PYTHONPATH=src python main.py --mode demo

# Single query
source .venv/bin/activate
PYTHONPATH=src python main.py --query "your query here"

✅ Done!

Your modern AWS DevOps Agent is ready with:

  • ✅ Modern Python structure (src/ layout, pyproject.toml)
  • ✅ Real AWS MCP integration (Cost Explorer, CloudWatch, Pricing)
  • ✅ Organized tools (cost, IaC, compliance, GitHub, CDK)
  • ✅ CDK Analysis (project analysis, synthesis, optimization)
  • ✅ Production deployment (Bedrock Agent Core)
  • ✅ Development tools (pytest, black, mypy)

📖 Next Steps

  • Learn how to use it: See docs/APP_USAGE.md for comprehensive examples
  • Explore capabilities: Run make dev to see demo scenarios
  • Start optimizing: Run make run and ask about your AWS costs

🚀 Available Commands

Makefile Commands (Simplified)

# Main Commands (PRIORITIZED)
make run               # Start interactive mode - MAIN COMMAND
make dev               # Start demo mode
make query QUERY="your query"  # Single query

# Setup & Installation
make setup             # Complete setup: create venv, install deps, install MCP servers
make install           # Alias for setup (same as make setup)
make clean             # Clean up temporary files and virtual environment
make status            # Show project status

# MCP Server Management
make mcp-check         # Check if MCP servers are installed (AWS + GitHub)
make mcp-install       # Install AWS MCP servers + GitHub MCP server
make mcp-test          # Test MCP server connections

# Development & Testing
make test              # Run core tests (unit + document generation)
make test-integration  # Run integration tests (requires GitHub config)
make format            # Format code

# Examples & Demos
make example TYPE=cost # Run cost analysis example
make example TYPE=iac  # Run IaC analysis example
make example TYPE=compliance # Run compliance check example
make example TYPE=github # Test GitHub connectivity

# AgentCore Deployment (Production)
make agentcore-configure # Configure AgentCore for deployment
make agentcore-deploy    # Deploy to Bedrock AgentCore (PRODUCTION)
make agentcore-test      # Test deployed AgentCore agent

Manual Commands (Advanced)

# Interactive mode (requires virtual environment)
source .venv/bin/activate
PYTHONPATH=src python main.py --mode interactive

# Demo scenarios
source .venv/bin/activate
PYTHONPATH=src python main.py --mode demo

# Single query
source .venv/bin/activate
PYTHONPATH=src python main.py --query "Analyze AWS costs for my infrastructure"

# Run without account selection
source .venv/bin/activate
PYTHONPATH=src python main.py --mode interactive --no-account-selection

🚀 AgentCore Deployment (Production)

Quick AgentCore Setup

# 1. Validate environment
make agentcore-validate

# 2. Test locally
make agentcore-test-local

# 3. Build Docker image
make agentcore-build

# 4. Deploy to production (with human verification)
make agentcore-deploy

AgentCore Monitoring

# Health checks
make agentcore-health       # Check health status
make agentcore-metrics      # Get metrics

# Monitoring
make agentcore-status       # Deployment status
make agentcore-logs         # View logs
make agentcore-monitor      # Performance monitoring

AgentCore Configuration Files

  • deployment/bedrock/app.py - Main application
  • deployment/bedrock/Dockerfile - Container config
  • deployment/bedrock/.bedrock_agentcore.yaml - AgentCore config
  • deployment/bedrock/iam-policy.json - IAM permissions
  • deployment/bedrock/env.example - Environment template

🔧 IAM Permissions Needed:

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "ce:GetCostAndUsage",
                "ce:GetForecast", 
                "cloudwatch:GetMetricData",
                "cloudwatch:DescribeAlarms",
                "logs:StartQuery",
                "logs:GetQueryResults",
                "pricing:GetProducts",
                "pricing:DescribeServices"
            ],
            "Resource": "*"
        }
    ]
}

🎯 What's New

  • No version numbers in filenames (maintainable)
  • Domain-organized tools (easy to navigate)
  • Modern Python packaging (industry standard)
  • Multiple entry points (development & production)
  • Comprehensive testing (unit, integration, fixtures)

That's it! 🎉