Blog

Keeping automations alive.

Practical guides and articles on the reliability of n8n and Make workflows, written by someone who runs them in production.

Articles

Real cases and methods.

AI agents

n8n AI agent: build it, then monitor it in production

Building an AI agent in n8n takes an hour. Keeping it reliable means watching four things: failures, loops, silence and cost.

· 2 min read
Comparison

n8n vs Make vs Zapier: which is most reliable in production?

Comparisons talk about price and integrations. This one compares n8n, Make and Zapier on what matters once you are live: errors, recovery, monitoring.

· 2 min read
n8n

Self-hosted n8n (Docker, VPS): the production checklist

Installing n8n with Docker on a VPS takes twenty minutes. Here are the twelve points to settle before running a client's automations on it.

· 2 min read
n8n

What is n8n? Definition, pricing, limits

n8n is a workflow automation tool you can host yourself. What it is for, what it really costs, and what it does not do by itself.

· 2 min read
n8n

n8n Error Workflow: what it catches, and the 5 failures it will never see

The n8n error workflow is essential, but it only reacts to failed executions. Here are the five failures that slip past it and how to cover them.

· 2 min read
Make

Make scenario deactivated by itself: causes and how to be warned

Make deactivates a scenario after repeated errors. Why, how to avoid it, and how to know that a scenario has stopped running.

· 2 min read
Troubleshooting

Webhook not receiving anything: 7 causes and how to spot it

A mute webhook produces no error in n8n or Make. The seven most common causes and the method to detect it in hours, not weeks.

· 2 min read
Method

Heartbeat monitoring: watching a script, a cron job or a Zap

A heartbeat reverses the logic of alerting: you are warned when a message does not arrive. Principle, setup and limits.

· 2 min read
Agencies

Automation maintenance contracts: what to include and how to justify them

What an automation maintenance contract should cover, what it should exclude, and how to prove to your client every month that it is worth the money.

· 2 min read
Method

Uptime probe or workflow monitoring: why a ping is not enough

An uptime probe checks that the server answers. It says nothing about what your workflows are doing. What each approach covers, and when you need both.

· 2 min read

Your first monitored workflow in ten minutes.

Connect an n8n instance or a Make team: Nokto discovers your workflows, learns their rhythm and warns you at the first anomaly.

14 days free · no credit card · cancel anytime