This commit adds comprehensive support for JSON configuration files in Docker containers, addressing the issue where the Docker image fails to start in server mode and ignores configuration files. ## Changes ### Docker Configuration Support - Added parse-config.js to safely parse JSON configs and export as shell variables - Implemented secure shell quoting to prevent command injection - Added dangerous environment variable blocking for security - Support for all JSON data types with proper edge case handling ### Docker Server Mode Fix - Added support for "n8n-mcp serve" command in entrypoint - Properly transforms serve command to HTTP mode - Fixed missing n8n-mcp binary issue in Docker image ### Security Enhancements - POSIX-compliant shell quoting without eval - Blocked dangerous variables (PATH, LD_PRELOAD, etc.) - Sanitized configuration keys to prevent invalid shell variables - Protection against shell metacharacters in values ### Testing - Added 53 comprehensive tests for Docker configuration - Unit tests for parsing, security, and edge cases - Integration tests for Docker entrypoint behavior - Security-focused tests for injection prevention ### Documentation - Updated Docker README with config file mounting examples - Enhanced troubleshooting guide with config file issues - Added version bump to 2.8.2 ### Additional Files - Included deployment-engineer and technical-researcher agent files 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
12 KiB
Docker Deployment Guide for n8n-MCP
This guide provides comprehensive instructions for deploying n8n-MCP using Docker.
🚀 Quick Start
Prerequisites
- Docker Engine 20.10+ (Docker Desktop on Windows/macOS, or Docker Engine on Linux)
- Docker Compose V2
- (Optional) openssl for generating auth tokens
1. HTTP Server Mode (Recommended)
The simplest way to deploy n8n-MCP is using Docker Compose with HTTP mode:
# Clone the repository
git clone https://github.com/czlonkowski/n8n-mcp.git
cd n8n-mcp
# Create .env file with auth token
cat > .env << EOF
AUTH_TOKEN=$(openssl rand -base64 32)
USE_FIXED_HTTP=true
EOF
# Start the server
docker compose up -d
# Check logs
docker compose logs -f
# Test the health endpoint
curl http://localhost:3000/health
2. Using Pre-built Images
Pre-built images are available on GitHub Container Registry:
# Pull the latest image (~280MB optimized)
docker pull ghcr.io/czlonkowski/n8n-mcp:latest
# Run with HTTP mode
docker run -d \
--name n8n-mcp \
-e MCP_MODE=http \
-e USE_FIXED_HTTP=true \
-e AUTH_TOKEN=your-secure-token \
-p 3000:3000 \
ghcr.io/czlonkowski/n8n-mcp:latest
📋 Configuration Options
Environment Variables
| Variable | Description | Default | Required |
|---|---|---|---|
MCP_MODE |
Server mode: stdio or http |
stdio |
No |
AUTH_TOKEN |
Bearer token for HTTP authentication | - | Yes (HTTP mode)* |
AUTH_TOKEN_FILE |
Path to file containing auth token (v2.7.5+) | - | Yes (HTTP mode)* |
PORT |
HTTP server port | 3000 |
No |
NODE_ENV |
Environment: development or production |
production |
No |
LOG_LEVEL |
Logging level: debug, info, warn, error |
info |
No |
NODE_DB_PATH |
Custom database path (v2.7.16+) | /app/data/nodes.db |
No |
*Either AUTH_TOKEN or AUTH_TOKEN_FILE must be set for HTTP mode. If both are set, AUTH_TOKEN takes precedence.
Configuration File Support (v2.8.2+)
You can mount a JSON configuration file to set environment variables:
# Create config file
cat > config.json << EOF
{
"MCP_MODE": "http",
"AUTH_TOKEN": "your-secure-token",
"LOG_LEVEL": "info",
"N8N_API_URL": "https://your-n8n-instance.com",
"N8N_API_KEY": "your-api-key"
}
EOF
# Run with config file
docker run -d \
--name n8n-mcp \
-v $(pwd)/config.json:/app/config.json:ro \
-p 3000:3000 \
ghcr.io/czlonkowski/n8n-mcp:latest
The config file supports:
- All standard environment variables
- Nested objects (flattened with underscore separators)
- Arrays, booleans, numbers, and strings
- Secure handling with command injection prevention
- Dangerous variable blocking for security
Docker Compose Configuration
The default docker-compose.yml provides:
- Automatic restart on failure
- Named volume for data persistence
- Memory limits (512MB max, 256MB reserved)
- Health checks every 30 seconds
- Container labels for organization
Custom Configuration
Create a docker-compose.override.yml for local customizations:
# docker-compose.override.yml
services:
n8n-mcp:
ports:
- "8080:3000" # Use different port
environment:
LOG_LEVEL: debug
NODE_ENV: development
volumes:
- ./custom-data:/app/data # Use local directory
🔧 Usage Modes
HTTP Mode (Remote Access)
Perfect for cloud deployments and remote access:
# Start in HTTP mode
docker run -d \
--name n8n-mcp-http \
-e MCP_MODE=http \
-e AUTH_TOKEN=your-secure-token \
-p 3000:3000 \
ghcr.io/czlonkowski/n8n-mcp:latest
Configure Claude Desktop with mcp-remote:
{
"mcpServers": {
"n8n-remote": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/mcp-remote@latest",
"connect",
"http://your-server:3000/mcp"
],
"env": {
"MCP_AUTH_TOKEN": "your-secure-token"
}
}
}
}
Stdio Mode (Local Direct Access)
For local Claude Desktop integration without HTTP:
# Run in stdio mode (interactive)
docker run --rm -i --init \
-e MCP_MODE=stdio \
-v n8n-mcp-data:/app/data \
ghcr.io/czlonkowski/n8n-mcp:latest
Server Mode (Command Line)
You can also use the serve command to start in HTTP mode:
# Using the serve command (v2.8.2+)
docker run -d \
--name n8n-mcp \
-e AUTH_TOKEN=your-secure-token \
-p 3000:3000 \
ghcr.io/czlonkowski/n8n-mcp:latest serve
Configure Claude Desktop:
{
"mcpServers": {
"n8n-docker": {
"command": "docker",
"args": [
"run",
"--rm",
"-i",
"--init",
"-e", "MCP_MODE=stdio",
"-v", "n8n-mcp-data:/app/data",
"ghcr.io/czlonkowski/n8n-mcp:latest"
]
}
}
}
🏗️ Building from Source
Build Locally
# Clone repository
git clone https://github.com/czlonkowski/n8n-mcp.git
cd n8n-mcp
# Build image
docker build -t n8n-mcp:local .
# Run your local build
docker run -d \
--name n8n-mcp-local \
-e MCP_MODE=http \
-e AUTH_TOKEN=test-token \
-p 3000:3000 \
n8n-mcp:local
Multi-architecture Build
Build for multiple platforms:
# Enable buildx
docker buildx create --use
# Build for amd64 and arm64
docker buildx build \
--platform linux/amd64,linux/arm64 \
-t n8n-mcp:multiarch \
--load \
.
🔍 Health Monitoring
Health Check Endpoint
The container includes a health check that runs every 30 seconds:
# Check health status
curl http://localhost:3000/health
Response example:
{
"status": "healthy",
"uptime": 120.5,
"memory": {
"used": "8.5 MB",
"rss": "45.2 MB",
"external": "1.2 MB"
},
"version": "2.3.0",
"mode": "http",
"database": {
"adapter": "better-sqlite3",
"ready": true
}
}
Docker Health Status
# Check container health
docker ps --format "table {{.Names}}\t{{.Status}}"
# View health check logs
docker inspect n8n-mcp | jq '.[0].State.Health'
🔒 Security Considerations
Authentication
n8n-MCP supports two authentication methods for HTTP mode:
Method 1: AUTH_TOKEN (Environment Variable)
- Set the token directly as an environment variable
- Simple and straightforward for basic deployments
- Always use a strong token (minimum 32 characters)
# Generate secure token
openssl rand -base64 32
# Use in Docker
docker run -e AUTH_TOKEN=your-secure-token ...
Method 2: AUTH_TOKEN_FILE (File Path) - NEW in v2.7.5
- Read token from a file (Docker secrets compatible)
- More secure for production deployments
- Prevents token exposure in process lists
# Create token file
echo "your-secure-token" > /path/to/token.txt
# Use with Docker secrets
docker run -e AUTH_TOKEN_FILE=/run/secrets/auth_token ...
Best Practices
- Never commit tokens to version control
- Rotate tokens regularly
- Use AUTH_TOKEN_FILE with Docker secrets for production
- Ensure token files have restricted permissions (600)
Network Security
For production deployments:
- Use HTTPS - Put a reverse proxy (nginx, Caddy) in front
- Firewall - Restrict access to trusted IPs only
- VPN - Consider VPN access for internal use
Example with Caddy:
your-domain.com {
reverse_proxy n8n-mcp:3000
basicauth * {
admin $2a$14$... # bcrypt hash
}
}
Container Security
- Runs as non-root user (uid 1001)
- Read-only root filesystem compatible
- No unnecessary packages installed
- Regular security updates via GitHub Actions
📊 Resource Management
Memory Limits
Default limits in docker-compose.yml:
- Maximum: 512MB
- Reserved: 256MB
Adjust based on your needs:
services:
n8n-mcp:
deploy:
resources:
limits:
memory: 1G
reservations:
memory: 512M
Volume Management
# List volumes
docker volume ls | grep n8n-mcp
# Inspect volume
docker volume inspect n8n-mcp-data
# Backup data
docker run --rm \
-v n8n-mcp-data:/source:ro \
-v $(pwd):/backup \
alpine tar czf /backup/n8n-mcp-backup.tar.gz -C /source .
# Restore data
docker run --rm \
-v n8n-mcp-data:/target \
-v $(pwd):/backup:ro \
alpine tar xzf /backup/n8n-mcp-backup.tar.gz -C /target
Custom Database Path (v2.7.16+)
You can specify a custom database location using NODE_DB_PATH:
# Use custom path within mounted volume
docker run -d \
--name n8n-mcp \
-e MCP_MODE=http \
-e AUTH_TOKEN=your-token \
-e NODE_DB_PATH=/app/data/custom/my-nodes.db \
-v n8n-mcp-data:/app/data \
-p 3000:3000 \
ghcr.io/czlonkowski/n8n-mcp:latest
Important Notes:
- The path must end with
.db - For data persistence, ensure the path is within a mounted volume
- Paths outside mounted volumes will be lost on container restart
- The directory will be created automatically if it doesn't exist
🐛 Troubleshooting
Common Issues
Container Exits Immediately
# Check logs
docker logs n8n-mcp
# Common causes:
# - Missing AUTH_TOKEN in HTTP mode
# - Database initialization failure
# - Port already in use
Database Not Initialized
# Manually initialize database
docker exec n8n-mcp node dist/scripts/rebuild.js
# Or recreate container with fresh volume
docker compose down -v
docker compose up -d
Permission Errors
# Fix volume permissions
docker exec n8n-mcp chown -R nodejs:nodejs /app/data
Debug Mode
Enable debug logging:
docker run -d \
--name n8n-mcp-debug \
-e MCP_MODE=http \
-e AUTH_TOKEN=test \
-e LOG_LEVEL=debug \
-p 3000:3000 \
ghcr.io/czlonkowski/n8n-mcp:latest
Container Shell Access
# Access running container
docker exec -it n8n-mcp sh
# Run as root for debugging
docker exec -it -u root n8n-mcp sh
🚀 Production Deployment
Recommended Setup
- Use Docker Compose for easier management
- Enable HTTPS with reverse proxy
- Set up monitoring (Prometheus, Grafana)
- Configure backups for the data volume
- Use secrets management for AUTH_TOKEN
Example Production Stack
# docker-compose.prod.yml
services:
n8n-mcp:
image: ghcr.io/czlonkowski/n8n-mcp:latest
restart: always
environment:
MCP_MODE: http
AUTH_TOKEN_FILE: /run/secrets/auth_token
NODE_ENV: production
secrets:
- auth_token
networks:
- internal
deploy:
resources:
limits:
memory: 1G
reservations:
memory: 512M
nginx:
image: nginx:alpine
restart: always
ports:
- "443:443"
volumes:
- ./nginx.conf:/etc/nginx/nginx.conf:ro
- ./certs:/etc/nginx/certs:ro
networks:
- internal
- external
networks:
internal:
external:
secrets:
auth_token:
file: ./secrets/auth_token.txt
📦 Available Images
ghcr.io/czlonkowski/n8n-mcp:latest- Latest stable releaseghcr.io/czlonkowski/n8n-mcp:2.3.0- Specific versionghcr.io/czlonkowski/n8n-mcp:main-abc123- Development builds
Image Details
- Base:
node:22-alpine - Size: ~280MB compressed
- Features: Pre-built database with all node information
- Database: Complete SQLite with 525+ nodes
- Architectures:
linux/amd64,linux/arm64 - Updated: Automatically via GitHub Actions
🔄 Updates and Maintenance
Updating
# Pull latest image
docker compose pull
# Recreate container
docker compose up -d
# View update logs
docker compose logs -f
Automatic Updates (Watchtower)
# Add to docker-compose.yml
services:
watchtower:
image: containrrr/watchtower
volumes:
- /var/run/docker.sock:/var/run/docker.sock
command: --interval 86400 n8n-mcp
📚 Additional Resources
🤝 Support
- Issues: GitHub Issues
- Discussions: GitHub Discussions
Last updated: July 2025 - Docker implementation v1.1