Self-hosted n8n support

Self-hosted n8n setup, maintenance and support.

Set up self-hosted n8n on your server with backups, upgrades, monitoring and a named owner. Monthly support retainer for workflows, failures and n8n version upgrades.

Start with your customers

What do people need to know before they choose you?

Your customers need clear answers and an easy next step. We review the information they see, fix the gaps, and check what changes.

From a question to an enquiry

  • 01Server setup
  • 02Backups and restore test
  • 03Upgrades
  • 04Monitoring and alerts
  • 05Failed-run triage
01

Why self-hosted n8n needs an owner.

Running n8n on your own server keeps workflow data on infrastructure you control and avoids per-execution pricing, but it also makes you responsible for everything n8n Cloud would otherwise handle: the server, the database, backups, security updates, n8n version upgrades, and noticing when a workflow fails at 2 a.m. Most self-hosted problems are not exotic. They are an encryption key that was never backed up, a database that filled the disk, an upgrade that changed a node's behaviour, or failed executions nobody looked at for a week. The support retainer exists so each of those has a named owner and a routine.

02

Setup: what a production-ready instance includes.

One-off setup is a fixed $499 (USD, excluding tax and server costs). It produces an n8n instance on your server or cloud account that you own, documented so another engineer could take it over.

  • n8n in Docker with PostgreSQL rather than the default SQLite
  • HTTPS behind a reverse proxy, with the editor protected and webhooks reachable
  • The encryption key stored safely outside the server, because credentials cannot be decrypted without it
  • Nightly database backups, plus a restore test before go-live
  • Execution data pruning so the database does not grow without limit
  • Uptime and failed-execution alerts sent to a named person
  • Queue mode with Redis and workers when volume needs it
  • A written runbook: how to restart, restore, upgrade and rotate credentials
03

The monthly support retainer.

Support is $1,349 per month (USD, excluding tax) and covers the routine work that keeps a self-hosted instance healthy, plus 10 hours a month for workflow changes and fixes.

  • Monthly n8n and server security updates, tested on a copy first
  • n8n version upgrades with release-note review for breaking changes
  • Backup checks and a quarterly restore test
  • Failed-execution review and fixes
  • Credential expiry and rotation tracking
  • Workflow changes: 10 hours per month included
  • Response time for production incidents agreed in the scope
04

A worked example: an upgrade without surprises.

A team running n8n for lead intake, invoicing reminders and a weekly report wants to move to a newer n8n version. Under the retainer, we read the release notes for every version in between and list the nodes and behaviours that changed. We take a database backup and copy the instance, including its encryption key, to a staging server. We upgrade the copy, run each production workflow against test inputs, and compare the outputs with the current version. One workflow uses a node whose output format changed, so we adjust its expression on the copy and re-test. Only then do we upgrade production in an agreed window, watch the first executions, and keep the backup ready for an agreed period in case a rollback is needed. The team receives a short note listing what changed.

05

How we start.

Every support engagement starts with an audit of the current instance, whether we set it up or not.

  • 1. Access review: server, database, n8n owner account, DNS and credentials
  • 2. Health audit: version, database, backups, disk, failed executions, exposed endpoints
  • 3. Fix list: the issues found, ranked by risk, and a fixed quote to resolve them
  • 4. Routine: backups, alerts, update schedule and runbook agreed
  • 5. Monthly report: updates applied, failures fixed, hours used and recommendations
06

Tested recovery, not just backups.

Marsen's published n8n recovery lab tests what happens when a callback repeats, a destination write succeeds but its response is lost, and both processes are killed and restarted. The same habit applies to support: a backup only counts once it has been restored and the workflows have been seen to run.

07

Self-hosted or n8n Cloud?

n8n Cloud is the simpler choice when you do not have a technical owner and your data-location rules allow it; n8n maintains the infrastructure and upgrades. Self-hosting suits teams that need data on their own infrastructure, high execution volumes, custom nodes or network access to internal systems. n8n's self-hosted Community Edition is free to use under its Sustainable Use License for internal business purposes; some features require a paid licence. We will recommend Cloud on the first call if it fits better.

Clear from
the first conversation.

We agree the work, the cost, and how we’ll check progress together.

A little more clarity

Good questions.
Straight answers.

Talk to our team
How much does self-hosted n8n support cost?

Setup is a fixed $499. The retainer is $1,349 per month and includes updates, backups, monitoring and 10 hours a month for workflow changes. Prices are in US dollars, excluding tax and server costs.

Can you take over an n8n instance someone else set up?

Yes. We start with an access review and health audit, then quote a fixed price to fix what we find before the retainer begins.

Where will n8n be hosted?

On your server or cloud account, so you own the instance and its data. We can recommend a provider and size, or work with your existing infrastructure.

What happens if the encryption key is lost?

Stored credentials cannot be decrypted without it and would have to be re-entered. Setup stores the key safely outside the server, and the runbook says where.

Do you upgrade n8n automatically?

No. Upgrades are tested on a copy against your workflows first, then applied in an agreed window with a backup ready for rollback.

Do you support queue mode and multiple workers?

Yes, when volume needs it. Queue mode uses Redis and separate worker processes, and changes how executions are monitored, so it is planned rather than switched on by default.

A practical next step

Tell us what
you want to improve.

Book a call

You don’t need a technical plan to start a conversation.