Observability for n8n at Scale

You built your first five n8n workflows and they worked. So you built ten more. Then twenty. Now you have 30 workflows running in production, processing real data, feeding real systems - and you have no idea if they are all healthy. A workflow failed at 3am on Saturday. Nobody noticed until the finance team asked on Monday morning why their reconciliation report was empty.

This is the observability gap that every automation programme hits. The tools that build workflows do not, by default, monitor them. And most teams discover this gap the hard way.

The watchdog architecture

We built a 5-workflow monitoring system entirely within n8n - no external observability platform required. Each workflow handles one concern, and together they give you complete visibility over your entire automation estate.

Workflow 1: Error Handler

This uses n8n's native Error Trigger node. Any workflow that fails anywhere in your instance triggers this handler. It captures the execution context - which workflow, which node, what the error was, what data was being processed - and sends a formatted email alert. The alert includes a direct link to the failed execution in n8n so the on-call person can click straight through to the problem.

Workflow 2: Execution Logger

Runs every five minutes. It queries the n8n API for all recent executions and logs them to a relational database with full metrics: execution duration, success or failure status, node count, start and finish timestamps. This gives you a historical record that survives n8n's built-in execution log pruning.

Workflow 3: Token Logger

If your workflows call AI models - and increasingly they do - you need to track token consumption. This workflow intercepts AI node outputs, logs the token count per model per execution, and calculates cost using your provider's pricing table. When a workflow that normally costs €0.02 per run suddenly costs €0.40, you get a flag.

Workflow 4: Weekly Git Sync

Every Sunday night, this workflow exports every workflow definition as JSON and commits the lot to a Git repository with a generated changelog. This is your safety net. If someone edits a production workflow and breaks it, you can diff against last week's version and see exactly what changed.

Workflow 5: Publish Sync

This triggers the moment anyone saves or activates a workflow. It immediately backs up the new version to both Git and the database. The weekly sync is your belt; this is your braces.

Self-monitoring and practical reality

• All telemetry is stored in the database, ready for dashboarding in whatever BI tool you already use.
• The entire system took two days to build and configure.
• It has run for months with zero maintenance.
• Adding new workflows to the monitored estate requires no configuration changes.

Automation without observability is just technical debt you have not discovered yet. Five workflows and a database table are all it takes to close the gap.