n8n's error workflow can't see a run that never happened

n8n's error workflow is a good safety net, but n8n's docs say it "runs if an execution fails." (n8n Docs) Two common problems aren't failures:

1. The run never starts. No execution means nothing fails, so nothing alerts.
2. The run goes green on empty data. Every node succeeds and nothing useful comes out.

The fix: expect a check-in, and get alerted when it's missing or empty.

• Add a last step. An HTTP Request node that POSTs the number of items processed, as items, to your heartbeat URL.
• Match the schedule. Give the watch your Schedule Trigger's schedule, plus a grace window for normal run time.
• Set a minimum item count of 1. A zero-item run alerts as empty. A run that never pings alerts as missed.

Set up an n8n watch, free for 3 watches.