An n8n node is failing silently and killing the whole run
A node hits a bad credential or a 5xx, throws, and the run just stops. No notification, no fallback — you only find out when a client asks where their data went.
Why it happens
The failable node has no error output wired, no 'Continue On Fail', and the workflow has no Error Workflow set, so any failure ends the run silently.
How to fix it
On each failable node enable 'Continue On Fail' or wire its error output to a handler branch, or set a workflow-level Error Workflow that alerts you.
How to verify
Break one node on purpose (bad credential) and confirm the run routes to your error path instead of stopping.
Not sure if this is the only leak?
Find every unguarded node in your workflow free — paste it and scan in 30 seconds.
Scan my workflow freeRelated leaks