// AUTOMATION FIXES
n8n workflow failed silently? Catch the runs that never happened
An n8n workflow failed silently when it stops doing its job without an error you’ll see: the trigger stopped firing, a credential expired, or a branch was skipped while the run still shows as completed. n8n’s Error Trigger only runs when an automatic workflow errors, so a workflow that never starts sends nothing. Web Blend runs its own silence alerts on n8n and builds n8n, Make.com and Zapier automations for trade and service businesses. We find the gap, then watch for silence.
Updated October 2026. n8n behaviour checked against n8n’s docs on 30 October 2026.
// CHECK IT NOW
Five checks in n8n
- When did it last execute? Open Executions for the workflow. A gap with no errors means the trigger stopped, not a node.
- Are there errors nobody read? Failed executions sit in the list, correctly marked, until someone looks. One builder on the n8n forum found days of them in their own instance.
- Is an error workflow set? In Workflow Settings, choose an Error workflow. Without one, failed runs alert no one.
- Did a credential need reconnecting? Google credentials from an app in “Testing” expire every 7 days. See connections that keep expiring.
- Do “successful” runs do the job? Open a recent one and check the last node wrote what it should. A skipped branch still shows as success.
// THE ERROR TRIGGER
The Error Trigger can only see runs that fail
The Error Trigger fires when an automatic run errors. It can’t fire for a run that never started, and you can’t test it with a manual run. A webhook stops being called, a polling trigger loses its credentials, an upstream app changes something: nothing runs, so nothing errors. Builders on the n8n community forum ask how to catch exactly this.
// HOW WE STOP IT
Watch the outcome, not just the errors
Prevention: our own account in your system
Connections run through a dedicated Web Blend user: Web Blend business name, an @webblend.au email, only the rights the automation needs. A staff password change or a deleted user can’t break it. You know it’s there and leave it alone.
Detection: an alert when a flow goes quiet
Error alerts need a run that fails. A trigger that stopped produces no run at all. So an n8n automation watches each flow for silence, for example no form leads in 3 days or no review requests sent, and alerts us.
Fail-safe: a register
Every action is logged to a Google Sheet, such as who was asked for a review. When something fails, the sheet shows exactly who missed out.
Recovery: a replay
A catch-up automation replays each missed action once, or backdates the data, such as conversions to Google Ads, so it catches up.
// PRICING
Automation Rescue + Watch
$490 + GST to fix what’s broken and set up monitoring, then $49/month + GST. For anyone running n8n, Cloud or self-hosted, including workflows we didn’t build. We watch Make scenarios too. See automation monitoring for small business.
// FIT
When you can do this yourself
-
You’re comfortable in n8n and can build a heartbeat workflow of your own. - The workflow is internal and nobody outside the business relies on it.
Running Make or Zapier as well? See Make scenario stopped running and Zapier stopped working.
// QUESTIONS
Questions owners ask
Not sure why your workflow went quiet? Tell us what it should have done.
Ask usWhy didn’t my n8n error workflow fire?
It only runs when an automatic execution errors. If the trigger never fired, there was no execution to fail. Manual runs don’t trigger it either.
How do I get alerted when an n8n workflow stops running?
Watch for silence. A second workflow checks that each flow produced its result in a set window and alerts you when it didn’t. Error alerts alone won’t catch it.
Why does n8n show success when the workflow didn’t do anything?
An IF or filter can route a run down a branch that does nothing, and the run still completes without an error. Check the output of the last node, not just the run status.
Why does my Google credential in n8n expire every week?
Google gives 7-day refresh tokens to apps left in “Testing” with external users. Move the app out of Testing or use another connection method.
Do you work on self-hosted n8n?
Yes. Watch covers self-hosted n8n as well as n8n Cloud, including workflows someone else built. Make scenarios too.
How much does it cost to fix and watch an n8n workflow?
Automation Rescue + Watch is $490 + GST to fix what’s broken and set up monitoring, then $49/month + GST.
// GET HELP
Know when a workflow goes quiet
Tell us which workflow stopped and what it should have done. We’ll tell you what probably happened and give you a fixed price.
// AUTOMATION RESCUE + WATCH
Tell us what stopped
Say what went quiet and roughly when. A person at Web Blend replies within one business day with what probably broke and a fixed price to fix it and watch it.