Manta SDK Usage Guide¶
The Manta SDK provides comprehensive tools for developing, deploying, and managing distributed computing workloads. This guide covers all aspects of using the Manta platform through its various interfaces.
🖥️ Dashboard Interface
Visual management of clusters, nodes, and swarms through the web dashboard.
- ✓ Cluster creation and management
- ✓ Real-time monitoring
- ✓ Result visualization
⌨️ Command Line Interface
Powerful CLI for configuration, deployment, and monitoring operations.
- ✓ SDK configuration
- ✓ Cluster operations
- ✓ Development workflow
🐍 Python API
Programmatic control for notebooks and applications.
- ✓ Swarm development
- ✓ Task orchestration
- ✓ Result collection
Overview¶
The Manta SDK provides three complementary interfaces for interacting with the platform:
Web Dashboard - Visual interface for cluster management and monitoring
CLI (Command Line Interface) - Terminal-based tools for configuration and operations
Python API - Programmatic interface for algorithm development and automation
Each interface is designed for specific use cases while maintaining consistency across the platform.
Contents¶
Interface Guides
Getting Started¶
Choose your interface based on your workflow:
- For Visual Management
Start with the Dashboard Guide to create clusters and monitor experiments through the web interface.
- For Development Workflows
Use the CLI Reference Guide for configuration management and quick operations from the terminal.
- For Algorithm Development
Learn Swarm Development Guide to build distributed algorithms using the Swarm and Task APIs.
- For Automation and Integration
Explore the Python User API for programmatic control in notebooks and applications.
Common Workflows¶
- Initial Setup
Configure SDK with credentials:
manta sdk config init
Verify connection:
manta sdk status
List available clusters:
manta sdk cluster list
- Development Cycle
Create a Swarm class with your algorithm
Test locally with mock data
Deploy to cluster using UserAPI
Monitor execution through dashboard or CLI
Collect results programmatically
- Production Deployment
Package modules with dependencies
Configure resource requirements
Deploy with fault tolerance settings
Monitor and scale as needed
Next Steps¶
New Users: Start with Getting Started with Dashboard for visual introduction
Developers: Jump to Swarm Basics for algorithm development
DevOps: Review Configuration Management for infrastructure setup
Data Scientists: Explore Using the API in Jupyter Notebooks for Jupyter integration