DataDog MCP Server

A Model Context Protocol (MCP) server that provides AI assistants with direct access to DataDog's observability platform through a standardized interface.

🎯 Purpose

This server bridges the gap between Large Language Models (LLMs) and DataDog's comprehensive observability platform, enabling AI assistants to:

• Monitor Infrastructure: Query dashboards, metrics, and host status
• Manage Events: Create and retrieve events for incident tracking
• Analyze Data: Access logs, traces, and performance metrics
• Automate Operations: Interact with monitors, downtimes, and alerts

🔧 What is MCP?

The Model Context Protocol (MCP) is a standardized way for AI assistants to interact with external tools and data sources. Instead of each AI system building custom integrations, MCP provides a common interface that allows LLMs to:

• Execute tools with structured inputs and outputs
• Access real-time data from external systems
• Maintain context across multiple tool calls
• Provide consistent, reliable integrations

📊 DataDog Platform

DataDog is a leading observability platform that provides:

• Infrastructure Monitoring: Track server performance, resource usage, and health
• Application Performance Monitoring (APM): Monitor application performance and user experience
• Log Management: Centralized logging with powerful search and analysis
• Real User Monitoring (RUM): Track user interactions and frontend performance
• Security Monitoring: Detect threats and vulnerabilities across your infrastructure

🚀 Quick Start

1. Build the server:

make build

2. Configure DataDog API:

export DD_API_KEY="your-datadog-api-key"
   export DATADOG_APP_KEY="your-datadog-app-key"  # Optional
   export DATADOG_SITE="datadoghq.eu"  # or datadoghq.com

3. Generate MCP configuration:

make create-mcp-config

4. Run the server:

./build/datadog-mcp-server

📚 Documentation

• **Available Tools** - Complete list of implementable DataDog tools
• **Test Documentation** - Test coverage and implementation details
• **OpenAPI Splitting** - How to split large OpenAPI specifications
• **Spectral Linting** - OpenAPI specification validation and linting
• **GitHub Actions** - CI/CD pipeline documentation

🛠️ Available Tools

Currently implemented tools include:

• Dashboard Management (v1): v1_list_dashboards, v1_get_dashboard
• Event Management (v1): v1_list_events, v1_create_event
• Connection Testing (v1): v1_test_connection
• Monitor Management (v1): (Coming soon)
• Metrics & Logs (v1): (Coming soon)

All tools are prefixed with their API version (e.g., v1_, v2_) for clear segregation and future v2 API support.

See docs/tools.md for the complete list and implementation status.

🔧 Development

# Install development tools
make install-dev-tools

# Run tests
make test

# Generate API client
make generate

# Split OpenAPI specifications
make split

# Lint OpenAPI specifications
make lint-openapi

# Build and test
make build

OpenAPI Management

The project includes comprehensive tools for managing OpenAPI specifications:

• Split Specifications: Break down large OpenAPI files into smaller, manageable pieces
• Spectral Linting: Validate OpenAPI specifications with custom rules and best practices
• Code Generation: Generate Go client code from OpenAPI specifications
• Version Support: Separate handling for DataDog API v1 and v2

See OpenAPI Splitting Guide and Spectral Linting Guide for detailed usage.

📚 Resources

• Model Context Protocol Introduction (Stytch Blog)