Detect Empty/Malformed Workflow Outputs with Keptick Receipt

Overview

This workflow provides a reusable pattern for checking that your n8n workflow produces the expected number of valid completed records. It demonstrates how to count distinct IDs from an output, detect empty results, malformed data, duplicates, and optionally send a lightweight receipt to Keptick for monitoring. The demo uses fictional data so you can test all scenarios without any external service — but you can easily adapt it to inspect real outputs from any node.

How It Works

The workflow is intentionally self-contained and driven by a configuration node. By changing a single parameter (demoCase), you can simulate healthy, empty, invalid, or duplicate results and see how the check behaves. The core logic lives in two Code nodes and one IF node; the HTTP request branch is optional and only runs when sendReceipt is enabled.

Node List

• Read first (Sticky Note) — Instructional note explaining the demo and how to adapt the workflow.
• Run the demo (Manual Trigger) — No auth. Starts the workflow on demand so you can test each scenario interactively.
• Configure check (Set Node) — No auth. Sets four configuration values: demoCase (healthy/empty/invalid/duplicates), minimumItems (minimum distinct IDs required), recordIdField (field name containing the record ID), and sendReceipt (boolean to enable Keptick delivery). Also includes testReceipt and receiptUrl for Keptick setup.
• Fictional completed outputs (Code Node) — No auth. Based on the demoCase value, returns a set of fictional records (or empty, malformed, or duplicate data). This node replaces the real output node in production.
• Check completed outputs (Code Node) — No auth. Extracts the configuration, filters out empty items, validates the recordIdField for each record, counts distinct IDs, tallies invalid and duplicate items, and compares against minimumItems. Returns a result object with check status, pass/fail, and a receipt payload.
• Send receipt enabled? (IF Node) — No auth. Evaluates sendReceipt from the configuration. If true, execution passes to the HTTP request branch; if false, execution goes to the local inspection node.
• Send count-only receipt (HTTP Request) — No explicit authentication. The receiptUrl contains the Keptick ingestion endpoint including a private token. Sends a POST request with the receipt payload (counts, status, execution ID). No retry on failure.
• Verify receipt accepted (Code Node) — No auth. Checks that the API response includes received: true. If not, throws an error. Returns an enriched result object.
• Inspect local result (NoOp) — No auth. A terminal node for viewing the check result directly in the editor when no receipt is sent.
• Use with a real workflow (Sticky Note) — Setup instructions for adapting the pattern to production.
• Optional Keptick monitor (Sticky Note) — Information about Keptick and how to obtain and configure the receipt URL.

Setup Instructions

1. No account required for demo: Run the workflow with demoCase: healthy (default) and inspect the output at the Inspect local result node. Change demoCase to empty, invalid, or duplicates to see how the guard handles each case.
2. Connect your real output: Replace the Fictional completed outputs node with the node that emits your completed records. Enable Always Output Data on that node so the check runs even when there are zero rows.
3. Configure the guard: Set recordIdField to the field name that holds your unique record identifier (e.g., id, orderId). Set minimumItems to the minimum number of distinct valid records you expect.
4. Optional Keptick monitoring: Create a free monitor at Keptick. Copy the private receipt URL and paste it into the receiptUrl field in Configure check. Set sendReceipt to true and testReceipt to true during setup. When ready for production, set testReceipt to false. Keep the receipt URL confidential — do not share workflow exports that contain it.
5. Test thoroughly: Run the workflow with healthy, empty, invalid, and duplicate data to verify your guard behaves as expected.

Use Cases & Adaptations

• Data pipeline validation: Ensure an ETL or sync workflow always produces a minimum number of records before triggering downstream steps.
• API batch processing: Check that all API calls returned valid IDs and no batches were empty or malformed.
• Duplicate detection: The guard counts distinct IDs, making it easy to spot duplicate outputs that might indicate replay or processing errors.
• Custom monitoring: Instead of Keptick, you can replace the HTTP request with a Slack message, email, or database write to alert on failures.
• Error handling: The guard only checks outputs after they are produced. Combine with n8n’s error workflow trigger to catch runs that never reach the check.

{
  "id": "keptickOutputGuard",
  "name": "Detect Empty Workflow Outputs with an Optional Keptick Receipt",
  "description": "Check distinct completed-record IDs, including empty and malformed output. Run a fictional local demo without an account; optionally send a count-only receipt to Keptick.",
  "active": false,
  "nodes": [
    {
      "id": "intro-note",
      "name": "Read first",
      "type": "n8n-nodes-base.stickyNote",
      "typeVersion": 1,
      "position": [
        -560,
        -340
      ],
      "parameters": {
        "content": "## Check the output, including zero\nRun the demo without an account or network request. In Configure check, change demoCase from healthy to empty, invalid, or duplicates. Inspect the last node.\n\nThe sample records are fictional. This checks distinct nonempty IDs in one field, not whether the real business work was correct.",
        "width": 570,
        "height": 245
      }
    },
    {
      "id": "trigger",
      "name": "Run the demo",
      "type": "n8n-nodes-base.manualTrigger",
      "typeVersion": 1,
      "position": [
        -540,
        40
      ],
      "parameters": {}
    },
    {
      "id": "config",
      "name": "Configure check",
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        -320,
        40
      ],
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "case",
              "name": "demoCase",
              "value": "healthy",
              "type": "string"
            },
// ... truncated (copy to see full JSON)