Self-host n8n with Managed Postgres

What does this n8n template deploy?

This template creates one n8n app and a separate Managed Postgres database. Lizard connects the app to Postgres through environment variable references, so you do not need to copy database passwords into a repository.

n8n lets you build workflows that connect APIs, transform data and run steps on a schedule or from a webhook. You can use it to route leads to a CRM, sync records between services, send alerts or call an AI model as part of a workflow. External services may charge for API calls, messages or model tokens.

The current template source wraps the official n8nio/n8n:latest image and listens on port 5678. It sets an app limit of 4 vCPUs and 4 GiB RAM, plus PostgreSQL 18 with 1 GiB of storage. These are template settings, not measured resource needs or a promise that every workflow will fit.

Postgres stores n8n's database records. The template does not set an explicit encryption key, attach Persistent Volumes to the app, or configure Redis workers. Complete the setup below before adding credentials or running business workflows.

How to deploy n8n on Lizard

1. Choose Deploy Now. Sign in to Lizard if needed. The template creates the app and database in your workspace. Check that your plan supports the resource limits shown above; the Free plan's limits are lower.
2. Check the deployment. Wait for the build and app to become healthy. Open the app's public HTTPS address and complete n8n's owner account setup. If startup fails, check the app logs and database connection variables.
3. Keep the encryption key stable. Before saving service credentials, set a strong, unique N8N_ENCRYPTION_KEY in the app's environment. Keep a secure backup of that key. For an existing instance, preserve its current key; replacing it with a new value does not re-encrypt existing credentials.
4. Set the public webhook address. Use your app's HTTPS URL in n8n's webhook configuration, then redeploy to apply environment changes. Test a published workflow through its production webhook URL.
5. Set the time zone and test a restart. Check scheduled triggers, an external API connection and saved workflow data after a controlled restart. Set up and test recovery before relying on the instance.

For repeatable service settings, read the Lizard CLI config reference. It covers lizard-config.json and lizard config apply --dry-run, which previews changes before you apply them. Keep secret values out of files you commit. For deployment or account help, contact Lizard support.

For a tested workflow, follow our n8n webhook to Postgres example. It includes a pinned n8n image, header authentication, retry checks and results from before and after a service restart.

How do webhooks work behind HTTPS?

n8n listens inside the app on port 5678; clients reach it through the public HTTPS address. Set the webhook base URL to that public address so n8n registers reachable callback URLs with other services.

For n8n 2.35.0 and later, use N8N_WEBHOOK_URL. Earlier versions use WEBHOOK_URL; n8n deprecated that name in 2.35.0. Match the setting to the version you deploy. Check N8N_PROXY_HOPS against the number of trusted proxies in your request path, and confirm forwarded host and protocol headers. See n8n's reverse proxy configuration.

After a domain change, update the base URL and any OAuth redirect URLs registered with external services. Test production webhooks separately from the editor's temporary test URLs.

Will workflows and credentials survive a redeploy?

The separate Postgres database stores workflow records and encrypted credentials. Recovering credentials also requires the original encryption key. n8n normally generates that key on first launch and saves it under ~/.n8n. A database backup alone cannot replace a lost key. Follow n8n's encryption key guidance.

Do not rely on the app's temporary filesystem for files you need to keep. If your setup needs local n8n files, attach a persistent volume at /home/node/.n8n and check that the container user can write to it. n8n's Docker setup guide explains why that directory still matters with Postgres.

Plan recovery for the database, encryption key and any stored files. Confirm backup retention and restore steps for your setup; this template does not configure a backup schedule. Read Lizard storage and recovery limits before choosing a recovery plan.

How much does self-hosting n8n cost?

Your total has three parts: Lizard hosting, any paid n8n license you choose, and charges from services your workflows call. n8n Community has no software subscription fee for permitted use, but that does not make the hosting or every n8n feature free.

Lizard offers billing only for active resources. For this template, budget for both the n8n app and Managed Postgres, plus storage and any billable egress. Include the plan fee and apply its credit once across the eligible usage in your account. A quiet n8n instance still uses memory and keeps its database running.

To estimate a month:

1. Measure the app's and database's CPU time and memory use over a representative period, including scheduled jobs and concurrent runs.
2. Multiply those quantities by the rates on Lizard pricing. Add storage and transfer charges after the applicable allowances.
3. Apply eligible plan credits without double-counting them. Add any n8n license and external API charges separately.

Execution count alone cannot predict the bill: a short HTTP request and a workflow that processes large files use different resources. The template's CPU and RAM limits are ceilings, not a fixed monthly charge. Measure your own workflows before choosing limits or comparing costs.

n8n Cloud, Lizard or a VPS: which should you choose?

Choose n8n Cloud when you want n8n to run the service. Choose Lizard when you want to self-host n8n with Managed Postgres and manage your app's settings without setting up a server. A VPS suits teams that want server-level control and can maintain that setup.

Community, Business and Enterprise do not include the same features. Check n8n's edition guide and current pricing for features such as sharing, SSO and environments. Review the n8n license for your intended use, especially if you plan to offer n8n to customers.

Can this template scale to more workflows?

Start by measuring run duration, memory peaks and concurrent executions. Tune workflow design and execution-data retention, then adjust app and database resources as needed. A larger resource limit alone does not make a slow API respond faster.

This template runs one n8n app. For a worker setup, n8n's queue mode adds Redis and worker processes that share the database and encryption key. It requires further configuration; adding app replicas alone is not the same setup. Lizard provides Managed Redis and Managed Postgres as separate services.

How should you update n8n?

The template currently follows the latest image tag, so a later build may use a newer n8n release. For controlled updates, use your own copy of the template repository and pin the image to a tested version. Back up the database and key, review n8n's release notes, and test important workflows before changing the running version. Test database restore as well as app rollback: a new release may change the database schema.