shelfio/datadog-mcp

Datadog MCP Server

A Model Context Protocol (MCP) server that provides comprehensive Datadog monitoring capabilities through Claude Desktop and other MCP clients.

Features

This MCP server enables Claude to:

• CI/CD Pipeline Management: List CI pipelines, extract fingerprints
• Service Logs Analysis: Retrieve and analyze service logs with environment and time filtering
• Metrics Monitoring: Query any Datadog metric with flexible filtering, aggregation, and field discovery
• Monitoring & Alerting: List and manage Datadog monitors and Service Level Objectives (SLOs)
• Service Definitions: List and retrieve detailed service definitions with metadata, ownership, and configuration
• Team Management: List teams, view member details, and manage team information

Quick Start

Choose your preferred method to run the Datadog MCP server:

🚀 UVX Direct Run (Recommended)

export DD_API_KEY="your-datadog-api-key" DD_APP_KEY="your-datadog-application-key"

# Latest version (HEAD)
uvx --from git+https://github.com/shelfio/datadog-mcp.git datadog-mcp

# Specific version (recommended for production)
uvx --from git+https://github.com/shelfio/datadog-mcp.git@v0.0.5 datadog-mcp

# Specific branch
uvx --from git+https://github.com/shelfio/datadog-mcp.git@main datadog-mcp

🔧 UV Quick Run (Development)

export DD_API_KEY="your-datadog-api-key" DD_APP_KEY="your-datadog-application-key"
git clone https://github.com/shelfio/datadog-mcp.git /tmp/datadog-mcp && cd /tmp/datadog-mcp && uv run ddmcp/server.py

🐳 Podman (Optional)

podman run -e DD_API_KEY="your-datadog-api-key" -e DD_APP_KEY="your-datadog-application-key" -i $(podman build -q https://github.com/shelfio/datadog-mcp.git)

Method Comparison:

Requirements

For UVX/UV Methods

• Python 3.13+
• UV package manager (includes uvx)
• Datadog API Key and Application Key

For Podman Method

• Podman
• Datadog API Key and Application Key

Version Management

When using UVX, you can specify exact versions for reproducible deployments:

Version Formats

• Latest: git+https://github.com/shelfio/datadog-mcp.git (HEAD)
• Specific Tag: git+https://github.com/shelfio/datadog-mcp.git@v0.0.5
• Branch: git+https://github.com/shelfio/datadog-mcp.git@main
• Commit Hash: git+https://github.com/shelfio/datadog-mcp.git@59f0c15

Recommendations

• Production: Use specific tags (e.g., @v0.0.5) for stability
• Development: Use latest or specific branch for newest features
• Testing: Use commit hashes for exact reproducibility

See GitHub releases for all available versions.

Claude Desktop Integration

Using UVX (Recommended)

Add to Claude Desktop configuration:

Latest version (bleeding edge):

{
  "mcpServers": {
    "datadog": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/shelfio/datadog-mcp.git", "datadog-mcp"],
      "env": {
        "DD_API_KEY": "your-datadog-api-key",
        "DD_APP_KEY": "your-datadog-application-key"
      }
    }
  }
}

Specific version (recommended for production):

{
  "mcpServers": {
    "datadog": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/shelfio/datadog-mcp.git@v0.0.5", "datadog-mcp"],
      "env": {
        "DD_API_KEY": "your-datadog-api-key",
        "DD_APP_KEY": "your-datadog-application-key"
      }
    }
  }
}

Using Local Development Setup

For development with local cloned repository:

git clone https://github.com/shelfio/datadog-mcp.git
cd datadog-mcp

Add to Claude Desktop configuration:

{
  "mcpServers": {
    "datadog": {
      "command": "uv",
      "args": ["run", "ddmcp/server.py"],
      "cwd": "/path/to/datadog-mcp",
      "env": {
        "DD_API_KEY": "your-datadog-api-key",
        "DD_APP_KEY": "your-datadog-application-key"
      }
    }
  }
}

Installation Options

UVX Installation (Recommended)

Install and run directly from GitHub without cloning:

export DD_API_KEY="your-datadog-api-key"
export DD_APP_KEY="your-datadog-application-key"

# Latest version
uvx --from git+https://github.com/shelfio/datadog-mcp.git datadog-mcp

# Specific version (recommended for production)
uvx --from git+https://github.com/shelfio/datadog-mcp.git@v0.0.5 datadog-mcp

Development Installation

For local development and testing:

1. Clone the repository:

   git clone https://github.com/shelfio/datadog-mcp.git
   cd datadog-mcp
   

2. Install dependencies:

   uv sync
   

3. Run the server:

   export DD_API_KEY="your-datadog-api-key"
   export DD_APP_KEY="your-datadog-application-key"
   uv run ddmcp/server.py
   

Podman Installation (Optional)

For containerized environments:

podman run -e DD_API_KEY="your-key" -e DD_APP_KEY="your-app-key" -i $(podman build -q https://github.com/shelfio/datadog-mcp.git)

Tools

The server provides these tools to Claude:

list_ci_pipelines

Lists all CI pipelines registered in Datadog with filtering options.

Arguments:

• repository (optional): Filter by repository name
• pipeline_name (optional): Filter by pipeline name
• format (optional): Output format - "table", "json", or "summary"

get_pipeline_fingerprints

Extracts pipeline fingerprints for use in Terraform service definitions.

Arguments:

• repository (optional): Filter by repository name
• pipeline_name (optional): Filter by pipeline name
• format (optional): Output format - "table", "json", or "summary"

list_metrics

Lists all available metrics from Datadog for metric discovery.

Arguments:

• filter (optional): Filter to search for metrics by tags (e.g., 'aws:', 'env:', 'service:web')
• limit (optional): Maximum number of metrics to return (default: 100, max: 10000)

get_metrics

Queries any Datadog metric with flexible filtering and aggregation.

Arguments:

• metric_name (required): The metric name to query (e.g., 'aws.apigateway.count', 'system.cpu.user')
• time_range (optional): "1h", "4h", "8h", "1d", "7d", "14d", "30d"
• aggregation (optional): "avg", "sum", "min", "max", "count"
• filters (optional): Dictionary of filters to apply (e.g., {'service': 'web', 'env': 'prod'})
• aggregation_by (optional): List of fields to group results by
• format (optional): "table", "summary", "json", "timeseries"

get_metric_fields

Retrieves all available fields (tags) for a specific metric.

Arguments:

• metric_name (required): The metric name to get fields for
• time_range (optional): "1h", "4h", "8h", "1d", "7d", "14d", "30d"

get_metric_field_values

Retrieves all values for a specific field of a metric.

Arguments:

• metric_name (required): The metric name
• field_name (required): The field name to get values for
• time_range (optional): "1h", "4h", "8h", "1d", "7d", "14d", "30d"

list_service_definitions

Lists all service definitions from Datadog with pagination and filtering.

Arguments:

• page_size (optional): Number of service definitions per page (default: 10, max: 100)
• page_number (optional): Page number for pagination (0-indexed, default: 0)
• schema_version (optional): Filter by schema version (e.g., 'v2', 'v2.1', 'v2.2')
• format (optional): Output format - "table", "json", or "summary"

get_service_definition

Retrieves the definition of a specific service with detailed metadata.

Arguments:

• service_name (required): Name of the service to retrieve
• schema_version (optional): Schema version to retrieve (default: "v2.2", options: "v1", "v2", "v2.1", "v2.2")
• format (optional): Output format - "formatted", "json", or "yaml"

get_service_logs

Retrieves service logs with comprehensive filtering capabilities.

Arguments:

• service_name (required): Name of the service
• time_range (required): "1h", "4h", "8h", "1d", "7d", "14d", "30d"
• environment (optional): "prod", "staging", "backoffice"
• log_level (optional): "INFO", "ERROR", "WARN", "DEBUG"
• format (optional): "table", "text", "json", "summary"

list_monitors

Lists all Datadog monitors with comprehensive filtering options.

Arguments:

• name (optional): Filter monitors by name (substring match)
• tags (optional): Filter monitors by tags (e.g., 'env:prod,service:web')
• monitor_tags (optional): Filter monitors by monitor tags (e.g., 'team:backend')
• page_size (optional): Number of monitors per page (default: 50, max: 1000)
• page (optional): Page number (0-indexed, default: 0)
• format (optional): Output format - "table", "json", or "summary"

list_slos

Lists Service Level Objectives (SLOs) from Datadog with filtering capabilities.

Arguments:

• query (optional): Filter SLOs by name or description (substring match)
• tags (optional): Filter SLOs by tags (e.g., 'team:backend,env:prod')
• limit (optional): Maximum number of SLOs to return (default: 50, max: 1000)
• offset (optional): Number of SLOs to skip (default: 0)
• format (optional): Output format - "table", "json", or "summary"

get_teams

Lists teams and their members.

Arguments:

• team_name (optional): Filter by team name
• include_members (optional): Include member details (default: false)
• format (optional): "table", "json", "summary"

Examples

Ask Claude to help you with:

"Show me all CI pipelines for the shelf-api repository"

"Get error logs for the content service in the last 4 hours"

"List all available AWS metrics"

"What are the latest metrics for aws.apigateway.count grouped by account?"

"Get all available fields for the system.cpu.user metric"

"List all service definitions in my organization"

"Get the definition for the user-api service"

"List all teams and their members"

"Show all monitors for the web service"

"List SLOs with less than 99% uptime"

"Extract pipeline fingerprints for Terraform configuration"

Configuration

Environment Variables

Obtaining Datadog Credentials

1. Log in to your Datadog account
2. Go to Organization Settings → API Keys
3. Create or copy your API Key (this is your DD_API_KEY)
4. Go to Organization Settings → Application Keys
5. Create or copy your Application Key (this is your DD_APP_KEY)

Note: These are two different keys:

• API Key: Used for authentication with Datadog's API
• Application Key: Used for authorization and is tied to a specific user account