Self-host n8n on Ubuntu 24.04 for cross-border e-commerce

You run a storefront that sells into Vietnam but your warehouse sits in China and your supplier invoices in USD. Every morning you manually copy orders from Shopify into a spreadsheet, email the warehouse, then paste tracking numbers back. That workflow breaks the moment you get more than a handful of orders a day, and it breaks even faster when your automation tool lives in a datacenter on the other side of the planet. Self-hosting n8n on an Ubuntu 24.04 VPS in Vietnam fixes the latency problem at the source, because the server that receives the webhook is the same network your Vietnamese customers and partners already use.

Why run n8n on a Vietnam VPS for cross-border work

Cross-border e-commerce automation is mostly webhook traffic. Shopify fires an order event, your warehouse API confirms stock, your bank sends a payout notification. Every one of those events travels over the public internet, and if your n8n instance sits in Singapore or the US, each round trip adds 50 to 150 ms before you even start processing. That sounds small until a Chinese warehouse API times out at 5 seconds and your n8n workflow fails at step two.

The bigger issue is outbound IP reputation. When n8n calls a warehouse or a payment gateway, that service sees the IP address of your VPS. A clean, dedicated IPv4 in Vietnam that has never sent spam passes more API checks than a shared or recycled address. For webhook delivery specifically, the receiving service often checks the origin IP against blocklists. If your previous VPS provider gave you an IP that was used for scraping, you will chase authentication errors that have nothing to do with your code.

Hosting n8n inside Vietnam also matters when your workflow touches local systems: Vietnamese banks, domestic logistics carriers, the tax authority's APIs. Those endpoints frequently block or throttle foreign IPs. A Linux VPS with a Vietnamese IPv4 talks to them as a local service, not as an international interloper. This is the practical argument for keeping your automation close to the market you automate, and it is the reason the setup in this guide uses a Vietnam-based server.

在越南部署 n8n，用本地 IPv4 接收电商 webhook，比海外服务器更稳。

Running n8n in Vietnam with a local IPv4 for e-commerce webhooks is more stable than using an overseas server.

Prerequisites

• An Ubuntu 24.04 LTS VPS with at least 2 GB of RAM. n8n plus PostgreSQL and the n8n queue mode will fit, but 4 GB is the comfortable zone when you run several workflows.
• Root access or a user with sudo privileges.
• A domain name pointing to your VPS, because you will terminate TLS with Caddy or Nginx rather than exposing n8n on a raw port.
• Docker and Docker Compose v2 installed. If they are not set up yet, follow the standard Ubuntu install for docker-ce and the compose plugin.

Do not run n8n as a plain node process on a public port. The editor has no built-in authentication gate that is strong enough for the internet. You will put it behind a reverse proxy with TLS and lock the editor with n8n's own user management.

Step 1, Set up the base system and Docker

Start from a fresh Ubuntu 24.04 install. Update the package index, then install the packages you need for Docker's official repository:

sudo apt update
sudo apt install -y ca-certificates curl gnupg
sudo install -m 0755 -d /etc/apt/keyrings
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /etc/apt/keyrings/docker.gpg
echo \
  "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] https://download.docker.com/linux/ubuntu \
  $(. /etc/os-release && echo "$VERSION_CODENAME") stable" | \
  sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
sudo apt update
sudo apt install -y docker-ce docker-ce-cli containerd.io docker-compose-plugin

The docker-compose-plugin package gives you the v2 docker compose command, which is what the rest of this guide uses. Verify the install:

sudo docker run --rm hello-world
docker compose version

You should see the hello-world message and a compose version line. If the second command fails, log out and back in so your user is in the docker group, or run it with sudo.

Step 2, Write the Docker Compose file for n8n with PostgreSQL

Create a directory for n8n and a compose file inside it. This example uses PostgreSQL as the database instead of SQLite, which is the right call the moment you want the queue mode or more than one webhook running at the same time:

mkdir -p ~/n8n && cd ~/n8n
nano docker-compose.yml

services:
  postgres:
    image: postgres:16-alpine
    restart: unless-stopped
    environment:
      POSTGRES_USER: n8n
      POSTGRES_PASSWORD: change_this_password
      POSTGRES_DB: n8n
    volumes:
      - postgres_data:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U n8n"]
      interval: 10s
      timeout: 5s
      retries: 5

  n8n:
    image: docker.n8n.io/n8nio/n8n
    restart: unless-stopped
    ports:
      - "127.0.0.1:5678:5678"
    environment:
      DB_TYPE: postgresdb
      DB_POSTGRESDB_HOST: postgres
      DB_POSTGRESDB_USER: n8n
      DB_POSTGRESDB_PASSWORD: change_this_password
      DB_POSTGRESDB_DATABASE: n8n
      N8N_HOST: n8n.example.com
      N8N_PROTOCOL: https
      N8N_PORT: 5678
      WEBHOOK_URL: https://n8n.example.com/
      GENERIC_TIMEZONE: Asia/Ho_Chi_Minh
    volumes:
      - n8n_data:/home/node/.n8n
    depends_on:
      postgres:
        condition: service_healthy

volumes:
  postgres_data:
  n8n_data:

The n8n container listens on 127.0.0.1:5678 only, so it is never exposed directly to the internet. The WEBHOOK_URL variable is what n8n uses when it registers webhooks with Shopify or other services, and it must match your public domain or the registrations will point at the wrong host. Set GENERIC_TIMEZONE to Asia/Ho_Chi_Minh so scheduling nodes fire on Vietnam time.

Start the stack and confirm both containers are healthy:

docker compose up -d
docker compose ps

You want both postgres and n8n to show a healthy or running state. If the n8n container keeps restarting, check the logs with docker compose logs n8n; the usual cause is a password mismatch in the compose file.

Step 3, Put Caddy in front for automatic HTTPS

Exposing the editor even on a high port is a bad idea. Install Caddy on the host and let it terminate TLS and forward to the local n8n port. Caddy fetches and renews certificates automatically, which removes the certbot chore from your plate:

sudo apt install -y debian-keyring debian-archive-keyring apt-transport-https curl
curl -1sLf 'https://dl.cloudsmith.io/public/caddy/stable/gpg.key' | sudo gpg --dearmor -o /usr/share/keyrings/caddy-stable-archive-keyring.gpg
curl -1sLf 'https://dl.cloudsmith.io/public/caddy/stable/debian.deb.txt' | sudo tee /etc/apt/sources.list.d/caddy-stable.list
sudo apt update
sudo apt install caddy

Point your domain's A record at the VPS IPv4 before you start Caddy, otherwise the TLS handshake will fail. Then write a minimal Caddyfile:

sudo nano /etc/caddy/Caddyfile

n8n.example.com {
    reverse_proxy 127.0.0.1:5678
}

Reload Caddy and check that the certificate was issued:

sudo systemctl reload caddy
sudo journalctl -u caddy -n 20 --no-pager

You should see a log line about successfully obtaining a certificate. Open https://n8n.example.com in a browser and you will get the n8n setup screen, which is where you create the owner account.

Step 4, Open the firewall only for the ports you need

Ubuntu ships with ufw disabled by default. Enable it and allow only SSH, HTTP and HTTPS. Do not open port 5678, because Caddy is the only thing that talks to n8n and it does that over localhost:

sudo ufw allow OpenSSH
sudo ufw allow 80/tcp
sudo ufw allow 443/tcp
sudo ufw enable
sudo ufw status verbose

The status output should list the three rules and show the default policy as deny incoming. If you manage the VPS through a control panel that uses a non-standard SSH port, adjust the first rule to match it before you enable the firewall or you will lock yourself out.

Verify the whole path from the outside:

curl -I https://n8n.example.com

Expect an HTTP/2 200 header. If you get a timeout, the A record or the firewall rule is wrong; if you get a certificate error, the domain did not point here when Caddy started.

Step 5, Build a workflow for cross-border order sync

With n8n reachable over HTTPS, the real work starts. A typical cross-border flow for a Vietnam-facing storefront looks like this:

1. A Shopify Trigger node fires on orders/paid.
2. An HTTP Request node pushes the order to your Chinese warehouse API, mapping the Vietnamese shipping address into the fields the warehouse expects.
3. A Wait node polls for the warehouse response, because fulfilment takes minutes, not milliseconds.
4. An HTTP Request node sends the tracking number back to Shopify.
5. An Email Send node notifies your Vietnamese customer service inbox, which is a Gmail or Outlook account that talks to customers in Vietnamese.

Keep the language mapping in a separate step. Vietnamese addresses and customer names rarely match the fields a Chinese warehouse API was built for. Add a small Code node that normalises the payload before it leaves the server:

const order = $input.first().json;
return [{
  json: {
    order_id: order.name,
    recipient: order.shipping_address.name,
    phone: order.shipping_address.phone,
    province: order.shipping_address.province,
    district: order.shipping_address.city,
    address: `${order.shipping_address.address1} ${order.shipping_address.zip}`,
    items: order.line_items.map(item => ({
      sku: item.sku,
      qty: item.quantity
    }))
  }
}];

This is where a n8n VPS setup pays for itself. The workflow runs on a schedule or a webhook without you touching it, and a failed step retries according to the error workflow you attach. Test each node individually first, then activate the workflow and place a real test order.

Step 6, Back up n8n data before you need it

n8n stores credentials and workflow JSON in the n8n_data volume, and PostgreSQL holds the execution history. Losing the volume means rebuilding every credential by hand, which nobody wants to do at 2 AM. Back up both the Postgres database and the volume with a simple script run from cron:

mkdir -p ~/backups
sudo docker compose exec -T postgres pg_dump -U n8n n8n | gzip > ~/backups/n8n_$(date +\%F).sql.gz

That dumps the database. For the n8n files themselves, copy the named volume to a tar archive:

sudo tar czf ~/backups/n8n_volume_$(date +\%F).tar.gz -C /var/lib/docker/volumes/n8n_n8n_data/_data .

Schedule both with crontab -e and a daily entry, then copy the files off the server to object storage or another host. A backup that lives on the same disk as the VPS is not a backup, it is a recovery convenience.

If you rent the VPS from a provider that offers snapshots, take one before you upgrade the n8n image. VPS snapshots on KVM give you a rollback point that works even when the application-level backup is corrupted.

Troubleshooting common n8n failures

Webhooks from Shopify never arrive. Check that WEBHOOK_URL in the compose file matches your public domain exactly, including the trailing slash. Shopify validates the endpoint URL when you register it, and a mismatch silently drops the registration. Then confirm the port is reachable: curl -I https://n8n.example.com/webhook-test/<your-path> should return a 200 from n8n, not a 404 from Caddy.

n8n runs out of memory during a big import. A 2 GB VPS is workable for a handful of workflows, but queue mode with multiple workers is not. Check usage with docker stats and if the container sits above 1.5 GB, move to a 4 GB Ubuntu VPS or enable swap. Swap is a stopgap; it will not make a 2 GB box run 10 concurrent executions smoothly.

Certificate renewal fails after a few months. Caddy renews automatically, but only if port 80 stays open, because the HTTP challenge needs it. If you closed port 80 to "be more secure", renewal breaks. Keep 80 open and let Caddy redirect to HTTPS.

FAQ

Related articles

• Self-host n8n on an Ubuntu VPS with a Vietnam IPv4
• Deploy and secure n8n on a Debian VPS in Vietnam
• n8n on a dedicated IP, why webhook reliability depends on it
• How much RAM to run multiple n8n workflows