---
title: n8n
description: Run n8n workflow automation with Postgres and a volume, or in queue mode with a worker and Redis.
section: Templates
order: 3
updated: 2026-11-04
---

[n8n](https://n8n.io) automates work across your apps with a visual workflow editor and hundreds of integrations. The **n8n** template runs it from the official `n8nio/n8n:stable` image with its data in PostgreSQL.

## What the template creates

| Resource | What it is |
| --- | --- |
| `n8n` | The editor, API and webhooks on port 5678, with 512 MiB of memory and a 1 GB volume at `/home/node/.n8n` |
| `n8n-db` | PostgreSQL 17 with 1 GiB of storage, connected to `n8n` |
| `N8N_ENCRYPTION_KEY` | A shared variable Dekka generates, which encrypts the credentials you save in n8n |

The service's database settings come from its connection, and `WEBHOOK_URL` and `N8N_HOST` point at the service's public address, so webhooks you create in n8n show the right URL. The timezone is set to `Africa/Accra`, which you can change in the service's variables.

n8n needs nothing from you to deploy. Because the service needs 512 MiB and the database 1 GiB, it needs the Starter plan or above.

## After it deploys

Open the service's URL and create the owner account, which is the first thing n8n asks for. Then add credentials for the apps you want to automate and build your first workflow.

> [!WARNING]
> Keep `N8N_ENCRYPTION_KEY` if you ever move n8n elsewhere. Without it, n8n can't read the credentials saved in its database. You can reveal it from the shared variables on the **Environment** tab.

## Run n8n with workers

The **n8n with workers** template runs n8n in queue mode, where the main service hands each execution to a worker through Redis. It suits workflows that run often or for a long time, because the editor stays responsive while workers do the work.

| Resource | What it is |
| --- | --- |
| `n8n` | The editor, API and webhooks, as above, with `EXECUTIONS_MODE=queue` |
| `n8n-worker` | A worker running `n8n worker --concurrency=10`, with 512 MiB of memory |
| `n8n-db` | PostgreSQL 17, connected to both services |
| `n8n-queue` | Redis with 64 MiB and the `noeviction` policy, so queued jobs are never dropped |

Both services read the encryption key, execution mode and timezone from shared variables. To add workers later, duplicate `n8n-worker`, and each copy gets the same encryption key from the shared variable.

## Limits

The volume keeps n8n's local files, so the main service runs a single replica. During the pilot, deploys run on a simulated engine, so n8n doesn't actually start or serve its editor.

n8n is distributed under the [Sustainable Use License](https://github.com/n8n-io/n8n), which limits some commercial uses, so read its terms before you build on it.