SkynetLabs
Back to journal
AutomationNovember 17, 20266 min read

n8n Error Workflow: Get Alerted When an Automation Fails

By , Founder, SkynetLabs

How to get an alert when an n8n workflow fails: build an Error Trigger workflow, paste a copy-ready alert message, add retries, and test it the right way.

On this page

To get alerted when an n8n workflow fails, create a new workflow that starts with an Error Trigger node and ends in a Slack or email node. Publish it, then open each production workflow, go to Options, Settings, and pick it under Error workflow. One catch: it only fires when a published workflow fails on its own, not when you click Execute workflow in the editor, so a manual test will never trigger it.

By default a failed run just sits in the executions list until someone opens it, so a lead form can stop reaching your CRM on Tuesday and you find out on Friday. The steps below follow n8n's own docs for the Error Trigger node and its error handling guide, read on 17 November 2026. Menu labels can shift between versions, so check them against your own install. Any numbers below are examples, not measurements.

Why a failed run goes unnoticed

A production workflow runs without anyone watching. If a credential expires, an API changes a field name, or a spreadsheet column gets renamed, the affected node throws an error and the run stops. n8n records it, but nothing pushes that fact to you. For a small business with no one checking a dashboard every morning, a broken workflow can stay broken for days.

There are two separate problems here, and they need two separate fixes. A workflow that throws an error is solved with an error workflow, which is this post. A workflow that stops running without any error, such as a schedule that never fires, needs a different check. We covered that case in the silent lane watchdog.

Step by step: build the error workflow

  1. Create a new workflow and name it something obvious, for example "Error alerts".
  2. Add the Error Trigger node as the first node. This is the only trigger the workflow needs. It fires when another workflow fails and hands over details about the failure.
  3. Add a message node for the place you actually look. A Slack, Telegram, email or WhatsApp node all work. Pick the channel you check several times a day, not the one you check weekly.
  4. Build the message from the trigger data. The Error Trigger passes the workflow name, the execution ID and link, the error message and the last node that ran. Put those four things in the alert, and nothing else. A copy-paste version is in the next section.
  5. Save and publish the error workflow. n8n's docs say the workflow that contains the Error Trigger must be published. It does not need a schedule or a webhook of its own.
  6. Attach it to your real workflows. Open each production workflow, go to its settings, and choose "Error workflow" from the dropdown. Select "Error alerts". Repeat for every workflow you care about.
Testing trap: the error workflow only runs when the failing workflow was started automatically, by a trigger or a schedule. Clicking "Execute workflow" in the editor does not call it. To test, activate a throwaway workflow that is built to fail, such as a Code node that throws an error, and let its trigger fire.

What a useful alert says

An alert that says "something failed" wastes the first ten minutes of your morning. Write it so you can decide in five seconds whether it is urgent. Paste this into the message field of your Slack, email or Telegram node. The field names are the ones n8n lists for the Error Trigger output:

Workflow failed: {{ $json.workflow.name }}
Last node: {{ $json.execution.lastNodeExecuted }}
Error: {{ $json.execution.error.message }}
Open it: {{ $json.execution.url }}

A filled-in example, with made-up values:

LineExample
What brokeWorkflow: New enquiry to CRM
WhereLast node: Create contact
WhyError: 401 unauthorized
Where to lookLink to the execution

One exception from n8n's docs: if the failure is in the workflow's own trigger node, the run never starts, so there is no execution ID or URL, and the details arrive under trigger instead of execution. Use {{ $json.trigger.error.message }} for those cases, or add an IF node that picks the right message.

A 401 from a CRM node usually points to an expired or revoked credential, which is a two-minute fix once you know about it. The same fix discovered after four days means four days of enquiries that never reached the CRM.

If several workflows fail at once, for instance because one shared credential died, you will get one alert per workflow. That is fine for a handful of workflows. If it becomes noisy, send a single summary instead, but do not start filtering alerts out until you have lived with them for a few weeks.

Add retries so you only hear about real failures

Many failures fix themselves. An API times out, a service returns a brief server error, a rate limit clears after a few seconds. Alerting on every one of those trains you to ignore the channel. n8n has a node setting for this.

  • Retry On Fail is in each node's settings. Turn it on, set the maximum number of tries and the wait between them. As an example, three tries with a short wait between each covers most brief outages.
  • Be careful with retries on actions that send something. A retried email or text node can send twice if the first attempt actually went through before the error came back. Use retries freely on reads and lookups, and think twice on sends.
  • On Error controls what happens after the retries run out. The default stops the workflow, which triggers your error workflow. You can instead continue, or route the failed item down a separate error output, which suits batch jobs where one bad row should not stop the other hundred.

Raise your own errors for bad data

Some problems are not technical errors at all. A form arrives with no phone number, an invoice total comes through as zero, a date is in the past. The workflow runs happily and produces garbage. The Stop and Error node lets you fail on purpose with a message you write, so the same alert path catches it.

Put an IF node before the risky step, check the condition you care about, and send the bad branch into Stop and Error with a plain sentence such as "Enquiry has no contact method". The error workflow then reports it like any other failure, and you decide whether to fix the form or the workflow.

Make sure the alert itself cannot fail silently

The error workflow is a workflow, so it can break too. A few habits keep it honest:

  • Use a stable credential for the message node, not a personal login that gets reset when someone changes a password.
  • Send the alert to a shared channel, not one person's inbox, so a holiday does not become a blind spot.
  • Once a month, trigger a deliberate test failure and confirm the alert arrives.
  • Keep a list of your production workflows. Any new one needs the Error workflow setting chosen, and it is easy to forget on the one you built in a hurry.

How we handle this for clients

Every workflow we build for a client gets the error workflow attached before it goes live, with the alert going to a channel the client already uses. Failure handling is part of the build, not an add-on. If you want that set up on workflows you already run, or built from scratch, see our n8n automation service. If you are still deciding where n8n should run, read self-hosted n8n vs n8n Cloud first.

Questions people ask

Why is my n8n error workflow not triggering?

The usual cause is testing it by hand. The Error Trigger only runs when a workflow fails automatically, so clicking Execute workflow in the editor never calls it. Other causes: the error workflow is not published, or the failing workflow does not have it selected under Settings, Error workflow. Test with a published workflow built to fail and let its trigger fire.

Do I need a separate error workflow for each workflow?

No. One error workflow can be selected in the settings of many workflows. The alert includes the workflow name from the Error Trigger data, so you can tell which one failed.

Will n8n alert me if a workflow never runs at all?

No. The Error Trigger fires on a failed execution. A schedule that does not fire produces no execution, so it produces no error. That needs a separate check, covered in the silent lane watchdog post linked above.

Is this the same on n8n Cloud and self-hosted n8n?

The Error Trigger node and the Error workflow setting are part of n8n itself, so the steps are the same in both. Check the docs for your version if a menu label looks different.

Want this kind of result for your business?

Send a brief. Yes/no in 8 hours. No funnel.

Send a brief