Your n8n workflow succeeded. Did it do the work?

Choose a business result that the workflow can verify

A workflow can run without raising an error and still produce too little useful output. A search may return no matching records. A filter may remove everything. A destination may accept fewer writes than expected.

Choose a result close to the completed work: successfully created records, confirmed deliveries or files that were written and checked. Counting input rows is only appropriate if the outcome you care about is receiving those rows. Do not present the input count as proof that downstream work succeeded.

Send one receipt for the completed batch

Create a Keptick monitor, choose the expected interval and minimum output count, and copy its private endpoint. After the verification step, send a JSON POST like this:

{
  "eventId": "workflow-run-1042-final",
  "success": true,
  "items": 12,
  "durationMs": 3200
}

Use a unique event ID for each run. A retry of the same receipt should keep the same ID. Keptick deduplicates repeated IDs while its seven-day history is retained. Keep the endpoint private; possession of it allows outcome submission to that monitor.

The downloadable HTTP Request node runs once for its input batch. Replace its endpoint and place it after the node that represents verified completed output. It is a starting node, not an already connected monitoring system.

Handle zero output explicitly

In n8n, a downstream node may not execute when the preceding node produces no items. Attaching the receipt node to that output alone can therefore miss the very zero-result case you wanted to detect.

Create an explicit branch that emits a summary item when no results were produced, then send items: 0 from that branch. Do not enable an empty placeholder and then count that placeholder as one successful business output. If using “Always Output Data,” verify the actual completed-result count separately.

Keptick’s missing-run check can detect that no receipt arrived, but that is a different signal from an explicit zero-output receipt.

Cover execution errors and missing receipts

Import the separate error workflow, replace its endpoint and select it as the monitored workflow’s Error Workflow. Error workflows are intended for failed triggered executions; manual runs may behave differently. Test the actual scheduled or triggered path in a safe workflow.

Keptick evaluates an incoming outcome immediately. Missing-run checks start roughly every five minutes; a busy account queue can require another cycle, and email delivery can take longer. This is not an instant paging SLA.

Test three cases before depending on it

1. Normal output: the verified result meets the configured minimum and the workspace shows Healthy.
2. Empty output: a deliberate non-production zero-result run submits zero and shows Output below minimum.
3. No receipt: pause a test workflow long enough to pass its interval and grace period; confirm the missing-run state and intended alert route.

Also test a reported failure if you use an error workflow. Use fictional records or a separate test workflow so the check does not contact customers or alter production work.

One monitor and up to 500 outcomes per day are free. The paid workspace is $19/month, with up to 50 monitors and 15,000 outcomes per account per day. Existing tools such as Healthchecks.io and Cronitor may already cover your needs; start with the result or failure they are not currently showing you.