Running n8n in production: a week-one and month-six checklist

The DIY n8n vs managed comparison summarises who owns each responsibility. This guide goes one level deeper: what those responsibilities look like in practice, so you can judge whether your team can carry them.

Week one: getting it running

• Choose where n8n runs: a server you control or n8n Cloud
• Decide on the database: n8n uses SQLite by default, and PostgreSQL is the usual production choice and is required if you later move to queue mode
• Configure HTTPS, a domain, and secure admin access
• Store the encryption key that protects saved credentials somewhere safe; losing it means re-entering every credential
• Build the first workflow and test it against realistic, messy input

None of this is unusual for someone who has run web services before. For a team without that experience, it is where most DIY projects stall.

Month six: keeping it healthy

• Updating n8n and reading release notes for changes that affect your nodes
• Watching failed executions and fixing the cause, not just re-running them
• Handling a connected service that changes its API or revokes a token
• Checking that backups actually restore
• Adding new workflows without breaking the ones the business depends on

This is the part that is easy to underestimate. Automation that nobody watches tends to fail quietly, and the first sign is often a customer who never got a reply.

Skills a DIY setup needs

You do not need a large engineering team. You do need at least one person who is comfortable with APIs and JSON, can read error logs, understands basic server administration, and has time set aside for this work rather than squeezing it between other duties. If that person exists, DIY is a reasonable choice.

When managed makes more sense

• Nobody on the team can own hosting, credentials, and failures
• The workflows touch customers directly, so silent failures are costly
• You want AI steps, a custom CRM, or WhatsApp intake designed as one system
• You would rather approve requirements than debug executions

In a managed setup, Veyrox handles the setup, deployment, and ongoing maintenance of your n8n workflows, and your team approves requirements and handles escalated conversations. Details are on managed n8n automation.

A middle path: build managed, hand over later

Some businesses start managed to get the first production workflows right, then take them in-house once someone is trained. That works when the workflows are documented, exported, and kept in accounts the business can access from the start. Ask about handover on day one, not at the end of the engagement.

Questions to settle before you decide

• Who gets the alert at 11pm when a workflow fails?
• How many hours a month can someone realistically give this?
• Does any workflow handle data that must stay on your own servers?
• What would one day of a broken workflow cost in missed enquiries or manual work?

If hosting is still open, read self-hosted n8n vs n8n Cloud. If you are looking at hiring someone, read hiring an n8n developer in Egypt.