n8n Empty-Run Monitor and Workflow Watchdog

Find scheduled n8n workflows that finish successfully but return fewer items than you expect. Get a named finding, the last output count and a link to the execution, so you can investigate before the missing data reaches your next report.

See the result before connecting n8n

Start with empty input {}. The Actor returns five synthetic demo rows without contacting an n8n instance or charging a workflow event. Apify platform usage may still apply.

These examples come from the built-in demo, not customer incidents:

Workflow	Verdict	Last output	Why it needs attention
Nightly CRM sync	HEALTHY	118 items	Inside its daily window and above its minimum
Hourly price scrape	EMPTY	0 items	Expected at least 20 items
Daily invoice export	STALE	15 items	Last success was 2,610 minutes ago
Weekly cleanup	DARK	0 items	No successful execution available
Partner feed import	UNREADABLE	Unknown	The example API request was rejected

On a real run, each row also includes the workflow ID, final node, execution ID, timestamp and execution link. OUTPUT contains the totals and a compact list of findings for your alerting workflow.

Connect your instance

Enter your n8n base URL and an API key with access to read workflows and executions. Create the key in Settings > n8n API. The key field uses Apify secret-input encryption; credentials are not included in output rows.

Self-hosted n8n and n8n Cloud use the same public API. Your instance must retain successful execution data. If it does not, the Actor cannot inspect earlier output and will report the missing evidence.

With no watch list, the Actor scans active workflows with an enabled Schedule, Cron or Interval trigger. Webhook, manual and other on-demand workflows are excluded unless you select scanOnDemand.

For workflows with different schedules, enter their individual expectations:

{
  "n8nBaseUrl": "https://your-instance.example.com",
  "n8nApiKey": "YOUR_N8N_API_KEY",
  "watchList": [
    { "workflowId": "YOUR_DAILY_WORKFLOW_ID", "label": "Nightly CRM sync", "expectEveryMinutes": 1440, "minItems": 1 },
    { "workflowId": "YOUR_HOURLY_WORKFLOW_ID", "label": "Price import", "expectEveryMinutes": 60, "minItems": 20 }
  ]
}

Replace the placeholders before a live run. A workflow that legitimately returns no items can use minItems: 0; it will then be checked for freshness only.

Choose a schedule and know the cost

The Actor runs when you start or schedule it. A daily schedule is a useful starting point for daily jobs. More frequent checks increase the event count even when nothing has changed.

Each workflow checked costs $0.01 in Actor events. Prices below use a 30-day month and exclude any Apify platform usage:

Workload	Events	Actor event cost
Check 20 workflows once	20	$0.20
Check 20 workflows daily	600/month	$6.00/month
Check 20 workflows hourly	14,400/month	$144.00/month

An UNREADABLE result can still be a charged workflow check. Empty demo input creates no workflow charges. There are no browser or proxy requirements. See the current pricing details for platform usage.

What the verdict means

Each workflow gets one of five verdicts:

• HEALTHY: within its expected window and at or above its minimum item count.
• EMPTY: a recent successful execution returned fewer final-node items than the configured minimum. This can include a low, nonzero count.
• STALE: the latest success is older than the expected interval multiplied by graceMultiplier (default 1.5).
• DARK: no success is retained, or it is older than three stale windows.
• UNREADABLE: execution data was missing or inaccessible, so the workflow needs investigation.

From findings to an alert

Use an Apify integration or your own n8n workflow to read OUTPUT, filter the findings and send a Slack message, email or webhook. An alert destination is not configured automatically by running this Actor.

Pair it with the n8n Hygiene and Security Auditor when you also need to inspect how workflows are configured. This watchdog checks the recent execution evidence.

Limits to plan around

Only the latest successful execution is inspected. Failed runs need separate error monitoring. Counts come from the final executed node; they do not prove an email arrived or an external system accepted a change. Set your threshold to that node's actual output.

A missing execution can mean retention was disabled rather than that a workflow never ran. The default expectation is once a day, so use a watch list for hourly or weekly jobs. API permission errors must be resolved on your n8n instance.

Use it from an AI agent

The MCP tool scan_n8n_workflows accepts the same n8n connection and watch-list settings. It returns a summary, state counts and named findings directly. Omit the API key for a demo.

Endpoint: https://fractionalhqforyou--n8n-silent-success-auditor.apify.actor/mcp

You can also add fractionalhqforyou/n8n-silent-success-auditor to apify/actors-mcp-server. Live tool calls use the same workflow event price as batch runs.

Support and setup

I build and maintain this Actor. Use this Actor's Issues tab for a reproducible problem; include the run ID and expected result, never API keys.