One API blip and your n8n HTTP Request drops the lead
External APIs hiccup all the time. If your HTTP Request node has no retry, a single transient error means that lead is gone — no second attempt.
Why it happens
Retry On Fail is off (the n8n default), so the node gives up on the first failed call.
How to fix it
On each HTTP Request node turn on 'Retry On Fail' (e.g. 3 tries) and set a sensible timeout so slow calls don't hang the run.
How to verify
Point one call at a URL that returns 503 once and confirm it retries and still completes.
Not sure if this is the only leak?
Scan your workflow free to see which calls have no retry — 30 seconds.
Scan my workflow freeRelated leaks