Changing the default n8n port and adjusting network bindings lets you run multiple instances, secure your workflow server, and integrate smoothly with existing web proxies. By default, n8n listens on port 5678 and binds to all interfaces, which works well for local testing but requires modification for secure multi-user server environments.
This guide covers how to change ports using environment variables, configure Docker port mappings, bind specifically to local loopback addresses, and resolve common socket conflicts.
Changing the Default Port in n8n
The default port 5678 can be adjusted easily depending on how you deploy the platform. Whether you run n8n directly via Node.js, manage it as a systemd service, or containerize it, setting the port correctly prevents startup failures.
Environment Variables
The primary method for changing the port is the N8N_PORT environment variable. When starting n8n from the command line, prefix the command with the variable assignment:
N8N_PORT=8080 n8n start
If you manage n8n via a configuration file (~/.n8n/config, or specified via N8N_CONFIG_FILES), you can define the port parameter directly under the server block.
Systemd Service Setup
When running n8n as a background service on Linux via systemd, edit the service file to include the port environment variable under the [Service] section:
[Service]
Environment="N8N_PORT=8080"
ExecStart=/usr/bin/n8n start
After modifying the file, reload the daemon and restart the service:
sudo systemctl daemon-reload
sudo systemctl restart n8n
Configuring Network Binding and Host Interfaces
Port configuration goes hand-in-hand with host binding. By default, n8n binds to 0.0.0.0, meaning it accepts incoming connections from any network interface. If your server has multiple IP addresses or you want to restrict access to the local machine only, you must configure the host parameter.
Restricting to Localhost
To ensure the n8n dashboard is only accessible from the machine it runs on, set N8N_HOST to the loopback address:
N8N_HOST=127.0.0.1 N8N_PORT=5678 n8n start
This prevents external network requests from reaching the application directly, which is useful when putting the instance behind an SSH tunnel or a local reverse proxy.
Binding to Specific IP Addresses
In multi-tenant or multi-homed server environments, you might want n8n to listen exclusively on a specific internal network card. Set N8N_HOST to that exact interface IP:
N8N_HOST=192.168.1.50 n8n start
Port Mapping in Docker and Kubernetes
Containerized deployments decouple the internal application port from the port exposed on the host machine. This abstraction allows you to run multiple n8n containers simultaneously without port collisions.
Docker CLI Mapping
When starting a container, use the -p flag to map a host port to the default container port 5678. For example, to expose n8n on host port 9000:
docker run -d \
--name n8n \
-p 9000:5678 \
-v ~/.n8n:/home/node/.n8n \
n8nio/n8n
Inside the container, n8n still runs on port 5678, but external traffic hits port 9000 on the host.
Docker Compose Configuration
In a docker-compose.yml file, configure the ports block to map external ports to the internal 5678 port:
services:
n8n:
image: n8nio/n8n:latest
ports:
- "5679:5678"
environment:
- N8N_PORT=5678
For more details on container images and release tags, consult the official n8n documentation.
Reverse Proxy Integration
In production architectures, n8n sits behind a reverse proxy like Nginx or Traefik. The proxy handles SSL termination on port 443 and forwards traffic to the internal n8n port.
When setting up a reverse proxy, you must also configure the WEBHOOK_URL environment variable so n8n generates correct webhook callback links pointing to your public domain rather than the internal port.
WEBHOOK_URL=https://automation.yourdomain.com/
Failing to set this variable causes webhook triggers to fail because external services try to send data to internal ports like 5678 instead of your proxy endpoint.
Troubleshooting Port Conflicts
When n8n fails to start, error logs often point to networking issues. Here are common errors and fixes.
EADDRINUSE: Address Already in Use
This error occurs when another process or a lingering n8n instance occupies your chosen port.
- Identify the process using the port on Linux or macOS:
sudo lsof -i :5678 - Terminate the conflicting process or change your
N8N_PORTvalue to an unused port.
Permission Denied on Low Ports
Trying to bind n8n to ports below 1024 (such as port 80 or 443) without root privileges results in a permission error on Unix-based systems. Always use ports above 1024 for user-space applications, and let a reverse proxy handle standard web ports.
Frequently Asked Questions
Can I run multiple n8n instances on the same server?
Yes. Assign each instance a unique port using N8N_PORT and separate data directories using the N8N_USER_FOLDER environment variable to prevent database locks.
Why are my webhooks returning connection refused errors after changing ports?
If you change the port or move behind a reverse proxy, you must update the WEBHOOK_URL environment variable to reflect the new public-facing URL and port combination.
How do I check if my firewall is blocking the n8n port?
Use ufw on Ubuntu or check your cloud provider security groups. You can test local connectivity from another machine using nc -zv server_ip port_number.
Does changing the port affect my existing workflows?
No. Workflows are stored in the SQLite or PostgreSQL database. Changing the port only changes the network socket the server listens on, though any active browser tabs will need to reconnect to the new URL.
