amauo 3.0.5

Deploy Bacalhau compute nodes globally using AWS spot instances for cost-effective distributed computing

🌍 Amauo - AWS Spot Instance Deployment Tool

Modern AWS spot instance deployment tool for deploying Bacalhau compute nodes and sensor simulations. Features a clean Python package structure with beautiful Rich terminal output.

Deploy Bacalhau compute nodes across AWS regions using spot instances for cost-effective distributed computing.

🚀 One-Command Global Deployment

Deploy clusters across worldwide regions with a single command:

# Install and run instantly with uvx (no local setup required)
uvx amauo create

# Check status
uvx amauo list

# Setup configuration
uvx amauo setup

# Clean up
uvx amauo destroy

✨ PyPI Package Architecture

🌟 Zero Setup Required

• uvx execution: No installation or environment setup needed
• Python packaging: Clean CLI with Rich output and proper error handling
• Direct AWS deployment: Uses boto3 for native AWS integration
• Cross-platform: Works on any system with Python 3.12+

🔥 What This Provides

• ✅ Instant execution → uvx amauo works immediately
• ✅ Rich terminal UI → Beautiful tables and progress indicators
• ✅ Proper YAML parsing → Clean configuration management
• ✅ Type safety → Full type annotations throughout
• ✅ AWS-native → Direct boto3 integration, no container overhead

🎯 Superior Features

• Automatic spot preemption handling - Cost-effective deployment
• Built-in health monitoring - Comprehensive node validation
• Multi-region deployment - Distributed across AWS regions
• Deterministic node identity - Consistent sensor identities
• Secure credential management - Never commit secrets

🏗️ Architecture

Modern Stack

• Python Package: PyPI-distributed CLI with Rich output
• AWS Integration: Direct boto3 calls for native cloud operations
• Bacalhau: Distributed compute with Docker engine support
• YAML Processing: Proper PyYAML parsing and configuration management
• Type Safety: Full mypy type checking throughout

Deployment Flow

uvx amauo create → Python CLI → AWS boto3 → Spot Instance Deploy → Health check

Package Architecture

PyPI Package (amauo):
├── CLI framework with Rich UI
├── AWS Resource Manager
├── SSH Manager for remote operations
├── YAML configuration parsing
├── State management (JSON-based)
└── Node identity generation

📋 Prerequisites

Required

• Python 3.12+ (for uvx, usually already installed)
• AWS Account with EC2 permissions
• AWS Credentials configured in ~/.aws/credentials or environment

Automatic

The CLI automatically handles:

• AWS resource management (VPC, Security Groups, Key Pairs)
• Prerequisites checking (AWS access, SSH keys)
• YAML configuration parsing and validation
• File synchronization to remote nodes

🎛️ Configuration

Credential Setup

Create these files in the project directory before deployment:

# Required credential files
mkdir -p credentials/

# Bacalhau orchestrator endpoint
echo "nats://your-orchestrator.example.com:4222" > credentials/orchestrator_endpoint

# Bacalhau authentication token
echo "your-secret-token" > credentials/orchestrator_token

# AWS credentials for S3 access (optional)
cp ~/.aws/credentials credentials/aws-credentials

Deployment Settings

Edit config.yaml to customize:

aws:
  total_instances: 3
  username: ubuntu
  ssh_key_name: my-key
  files_directory: "deployment-files"
  scripts_directory: "instance/scripts"

regions:
  - us-west-2:
      machine_type: t3.medium
      image: auto  # Auto-discovers latest Ubuntu 24.04
  - us-east-1:
      machine_type: t3.medium
      image: auto

🔧 Commands

Core Operations

# Deploy instances across AWS regions
uvx amauo create

# List all running instances with details
uvx amauo list

# Destroy all instances
uvx amauo destroy

# Setup initial configuration
uvx amauo setup

# Show version
uvx amauo version

# Show help
uvx amauo help

# Migrate from legacy spot-deployer
uvx amauo migrate

Advanced Options

# Use custom config file (default: config.yaml)
uvx amauo create --config my-config.yaml

# Dry run to validate configuration without deploying
uvx amauo create --dry-run

# Verbose output for debugging
uvx amauo create --verbose

🧪 Local Development & Testing

Quick Test

# Test the CLI without installation
uvx amauo version

# Setup configuration
uvx amauo setup

# Test with dry run
uvx amauo create --dry-run

Local Development

# Clone the repository for development
git clone <repository-url>
cd bacalhau-skypilot

# Install in development mode with uv
uv pip install -e .

# Run locally during development
python -m amauo version

# Run tests
uv run pytest

# Run linting
uv run ruff check .

Debug Deployment

# Enable verbose logging for debugging
uvx amauo create --verbose

# Check instance status
uvx amauo list

# SSH to specific instance for debugging
# Use instance ID from the list command
ssh -i ~/.ssh/your-key ubuntu@instance-ip

Test Individual Components

# Test node identity generation
INSTANCE_ID=i-test123 python3 instance/scripts/generate_node_identity.py

# Test Bacalhau config generation
python3 instance/scripts/generate_bacalhau_config.sh

# Check deployment logs on instance
ssh -i ~/.ssh/your-key ubuntu@instance-ip sudo tail -f /opt/deployment.log

🌐 AWS Integration

Current Support

• AWS: Full native support with spot instances
• Multi-region: Deploy across multiple AWS regions simultaneously

Cloud Provider Detection

# AWS: Uses IMDS for instance metadata
curl -s http://169.254.169.254/latest/meta-data/instance-id

# Node identity automatically detects cloud provider
# and generates appropriate sensor identities

📊 Monitoring & Health

Built-in Health Checks

Every deployment includes comprehensive monitoring:

• Docker service status
• Container health (Bacalhau + Sensor)
• Network connectivity (API ports 1234, 4222)
• File system status (configs, data directories)
• Resource utilization (disk, memory)
• Orchestrator connectivity
• Log analysis (error detection)

Status Dashboard

# View instance overview
uvx amauo list

# SSH to specific node for debugging
ssh -i ~/.ssh/your-key ubuntu@instance-ip

# Check deployment logs
ssh -i ~/.ssh/your-key ubuntu@instance-ip sudo tail -f /opt/deployment.log

🔒 Security

Credential Management

• Never committed to git - credentials/ in .gitignore
• Secure file transfer via SSH to remote instances
• Encrypted in transit - SSH/TLS everywhere
• Least privilege - minimal required AWS permissions

Instance Security

• Official Ubuntu 24.04 LTS AMI - automatically discovered
• Security groups with minimal required ports
• SSH key-based access - no password authentication
• Automatic security updates via cloud-init

🚀 Performance

Deployment Speed

• ~3-5 minutes for multi-region deployment
• Parallel deployment across regions via boto3
• Fast startup time - ~0.15 seconds CLI response

Resource Efficiency

• t3.medium instances (2 vCPU, 4GB RAM) by default
• 30GB disk per node
• Spot pricing - up to 90% cost savings
• Efficient resource cleanup on destroy

Reliability

• Immutable infrastructure - destroy and recreate for changes
• Health monitoring with systemd services
• Multi-region distribution for availability
• AWS retry logic for transient API failures

🆘 Troubleshooting

Common Issues

1. AWS Credentials

# Check AWS access
aws sts get-caller-identity

# Configure if needed
aws configure
# or
aws sso login

2. SSH Key Issues

# Ensure SSH key exists and has correct permissions
chmod 400 ~/.ssh/your-key.pem

# Test SSH access to instance
ssh -i ~/.ssh/your-key.pem ubuntu@instance-ip

3. Configuration Issues

# Validate configuration file
uvx amauo create --dry-run

# Check current configuration
cat config.yaml

4. Instance Connectivity

# List current instances
uvx amauo list

# Check instance logs
ssh -i ~/.ssh/your-key ubuntu@instance-ip sudo tail -f /opt/deployment.log

Debug Commands

# Verbose deployment
uvx amauo create --verbose

# Check AWS resources directly
aws ec2 describe-instances --filters "Name=tag:amauo,Values=true"

# SSH to node for debugging
ssh -i ~/.ssh/your-key ubuntu@instance-ip
# Then run: sudo docker ps, sudo systemctl status bacalhau

🤝 Contributing

Development Setup

git clone <repository-url>
cd bacalhau-skypilot

# Install in development mode
uv pip install -e .

# Test the CLI
python -m amauo version

Testing Changes

1. Local testing: Use uvx amauo version to test CLI
2. Configuration test: Modify config.yaml and test parsing
3. Single node test: Deploy to one region first with minimal config
4. Full deployment test: Test complete multi-region deployment

Code Standards

• Python-first - native Python with boto3, no containers
• Type safety - full type annotations and mypy checking
• Rich UI - beautiful terminal output with progress indicators
• Immutable infrastructure - destroy and recreate for changes

📄 License

MIT License - see LICENSE for details.

🔗 Links

• Bacalhau Documentation: https://docs.bacalhau.org/
• AWS Documentation: https://docs.aws.amazon.com/
• uvx Documentation: https://docs.astral.sh/uv/guides/tools/

---

Ready to deploy? Ensure AWS credentials are configured, then: uvx amauo create