MCP Elasticsearch Server 🔍

MCP Elasticsearch Server 🔍

A Model Context Protocol (MCP) server that provides Elasticsearch integration capabilities to AI assistants and other MCP clients. This server enables querying Elasticsearch clusters through a standardized interface.

Features

• 🔐 Multiple Authentication Methods: Supports both API key and username/password authentication
• 📊 Index Management: List indices with health status and document counts
• 🗺️ Schema Discovery: Retrieve field mappings to understand index structure
• 🔍 Advanced Search: Execute complex Elasticsearch queries with aggregations and sorting
• 📋 Structured Responses: JSON-formatted output with search metadata
• ⚡ Performance Monitoring: Query execution time tracking
• 🎯 Context Aware: Supports search execution with proper context cancellation

Tools

list_indices

List all Elasticsearch indices with optional pattern filtering.

Parameters:

• pattern (string, optional): Index pattern filter (default: "*")

Returns:

• Total index count
• Index details (name, health, status, document count, size)

get_index_mappings

Get field mappings for one or more Elasticsearch indices.

Parameters:

• index (string, required): Index name or pattern

Returns:

• Complete field mappings for the specified indices

search

Execute Elasticsearch search queries with full DSL support.

Parameters:

• index (string, required): Index name or pattern to search
• query (string, optional): Elasticsearch query DSL as JSON (default: "{}")
• size (number, optional): Maximum documents to return (default: 10, max: 10000)
• sort (string, optional): Sort specification as JSON
• track_total_hits (boolean, optional): Track total hit count (default: true)

Returns:

• Search results with hits, aggregations, and metadata

Configuration

Environment Variables

Elasticsearch Configuration

• ES_URL: Elasticsearch cluster URL (required)
• ES_API_KEY: API key for authentication (optional)
• ES_USERNAME: Username for basic authentication (optional)
• ES_PASSWORD: Password for basic authentication (optional)

Server Configuration

• MCP_ES_SERVER_NAME: Server name (default: "mcp-elasticsearch 🔍")

Logging Configuration

• MCP_ES_LOG_LEVEL: Log level (debug, info, warn, error, fatal)
• MCP_ES_LOG_FORMAT: Log format (json, console)
• MCP_ES_LOG_OUTPUT: Log output (stdout, stderr)

Authentication

You must provide either:

1. API Key authentication: Set ES_API_KEY
2. Basic authentication: Set both ES_USERNAME and ES_PASSWORD

Installation

# Clone and build
git clone <repository>
cd mcp-elasticsearch
go mod download
go build -o bin/mcp-elasticsearch .

# Install to system
sudo install bin/mcp-elasticsearch /usr/local/bin/

Usage

Direct Execution

# With API key
ES_URL="https://your-cluster.com" ES_API_KEY="your-api-key" mcp-elasticsearch

# With basic auth
ES_URL="https://your-cluster.com" ES_USERNAME="user" ES_PASSWORD="pass" mcp-elasticsearch

# With custom logging
ES_URL="https://your-cluster.com" ES_API_KEY="key" MCP_ES_LOG_LEVEL=debug mcp-elasticsearch

Integration with Claude Desktop

Add to your Claude configuration:

{
  "mcpServers": {
    "elasticsearch": {
      "command": "mcp-elasticsearch",
      "env": {
        "ES_URL": "https://your-cluster.com",
        "ES_API_KEY": "your-api-key",
        "MCP_ES_LOG_LEVEL": "info"
      }
    }
  }
}

Integration with ApMentor

Update your config.json:

{
  "mcpServers": {
    "elasticsearch-go": {
      "command": "/usr/local/bin/mcp-elasticsearch",
      "env": {
        "ES_URL": "https://atani.es.eu-west-1.aws.found.io",
        "ES_API_KEY": "your-api-key",
        "MCP_ES_LOG_LEVEL": "info"
      }
    }
  }
}

Example Queries

List All Indices

{
  "tool": "list_indices",
  "parameters": {
    "pattern": "*"
  }
}

List Log Indices Only

{
  "tool": "list_indices", 
  "parameters": {
    "pattern": "logs-*"
  }
}

Get Index Mappings

{
  "tool": "get_index_mappings",
  "parameters": {
    "index": "logs-apm.error-*"
  }
}

Simple Search

{
  "tool": "search",
  "parameters": {
    "index": "logs-*",
    "query": "{\"match\": {\"service.name\": \"broker-api-b2b\"}}",
    "size": 50
  }
}

Complex Search with Aggregations

{
  "tool": "search",
  "parameters": {
    "index": "logs-*",
    "query": "{\"bool\": {\"must\": [{\"term\": {\"service.name\": \"broker-api-b2b\"}}, {\"range\": {\"@timestamp\": {\"gte\": \"now-24h\"}}}]}}",
    "size": 0,
    "aggs": "{\"error_types\": {\"terms\": {\"field\": \"error.type.keyword\", \"size\": 10}}}"
  }
}

Search with Sorting

{
  "tool": "search",
  "parameters": {
    "index": "logs-*",
    "query": "{\"match\": {\"log.level\": \"ERROR\"}}",
    "sort": "[{\"@timestamp\": {\"order\": \"desc\"}}]",
    "size": 20
  }
}

Development

# Install dependencies
go mod download

# Format code
gofmt -w .

# Run tests
go test -v ./...

# Build
go build -o bin/mcp-elasticsearch .

# Run with debug logging
MCP_ES_LOG_LEVEL=debug go run .

Error Handling

The server provides detailed error messages for common issues:

• Authentication failures: Check your API key or credentials
• Index not found: Verify index names and patterns
• Query syntax errors: Validate your Elasticsearch query JSON
• Connection issues: Ensure Elasticsearch is accessible

Security Considerations

• Store API keys and credentials securely
• Use environment variables for sensitive configuration
• Consider network security for Elasticsearch access
• Monitor query patterns and resource usage

License

MIT License - See LICENSE file for details.