agullon/microshift-mcp-server

MCP Server for MicroShift

A Model Context Protocol (MCP) server that exposes MicroShift management operations as standardized tools for AI models.

Features

• Smart Execution: Automatically detects local vs remote execution to optimize performance
• Service Management: Control MicroShift service (start, stop, restart, status)
• Log Retrieval: Get service logs for troubleshooting
• Pod Management: List and inspect cluster pods
• Configuration Management: View and update MicroShift configuration files
• Multi-Component Support: Handle microshift, lvmd, and ovn configurations

Quick Start

1. Set Environment Variables

export SSH_IP_ADDR="your-microshift-host"
export SSH_USER="microshift"
export SSH_CONFIG_FILE="~/.ssh/config"
export KUBECONFIG_PATH="/path/to/kubeconfig"

2. Start the Server

./start-mcp-server.sh

Available MCP Tools

• microshift_systemctl_commands(action) - Control MicroShift service
• get_latest_microshift_service_logs(number_of_lines) - Get service logs
• get_pods(namespace) - List cluster pods
• run_microshift_commands(action) - Run MicroShift commands
• get_microshift_current_config() - Get current MicroShift configuration
• get_default_config_yaml(component) - Get default configuration for microshift, lvmd, lvms, ovn
• get_microshift_observability_config() - Get current observability configuration
• get_microshift_custom_config() - Get custom configuration from config.d directory
• get_microshift_custom_manifests() - Get custom manifests from manifests.d and manifests directories
• override_microshift_config(component, override_config) - Update configuration for microshift, lvmd, lvms, ovn

Files

• mcp-server.py - Main server implementation using FastMCP
• start-mcp-server.sh - Startup script with auto-setup
• requirements.txt - Python dependencies
• mcp-server-vars.env - Environment configuration

Smart Execution

The server automatically detects if it's running on the same host as MicroShift:

• Local execution: Commands run directly for better performance
• Remote execution: Uses SSH when running on a different machine

Requirements

• Python 3.11+
• SSH access to MicroShift host (for remote operations)
• Valid kubeconfig file
• Sudo privileges on MicroShift host