Solution · reliability and meetings
Auto-tests for production automations: breakages are found before the start, not during it
The most dangerous thing in automation is a workflow that quietly broke the day before something important. This solution runs test data through the production pipelines in advance and on schedule, checks the result, sends a checklist to Telegram and cleans up after itself.
What it looks like
Mock-ups with demo data, no real data
Preflight
Pre-launch check · start in 3 days
✅ Questionnaire received
✅ Client record created
✅ Photo uploaded
✅ Team message sent
✅ QR code generated
· Test data removed
05:12
At a glance
- In production
- At a client under NDA
- Fits
- Anyone whose money, clients or events depend on automations
- Needs
- n8n automations with a checkable result
Estimate for your company
≈2–3 h
of manual checking before each important launch
Assumption: someone runs the form, payment and notifications by hand before the start (our assumption)
This is a calculation, not a promise: plug in your own volumes and the figure changes.
Before
- An external service changes its format or a token expires, and the automation silently stops working.
- The team learns about the breakage from clients, when it is too late to fix.
- Nobody has the time to check everything by hand before each launch.
What was built
Scheduled testing
Before each important launch — say, seven, three and one day ahead — the system runs the check itself. On quiet days it sends a short “system is alive” signal.
End-to-end run
A clearly labelled test questionnaire goes through the production pipeline, and the system checks that the record, page, message and QR code were created.
Checklist in Telegram
The result goes to a separate group: what passed and what did not.
Self-cleaning
Test pages are archived, and messages and spreadsheet rows are deleted, so no traces are left in working data.
An alert on any error
A global handler reports any production failure.
🧪Auto-test schedule
| When | What is checked | Result |
|---|---|---|
| 7 days before | Questionnaire, record, page, QR | All good |
| 3 days before | Same + attendance scan | All good |
| 1 day before | Final run | 1 error → fixing |
Stack
- n8n
- Notion
- Google Sheets
- Telegram
Where it runs
Questions
How is this different from error monitoring?
Monitoring catches a workflow that crashed. An auto-test catches a workflow that ran without errors but did not do its job.
Will test data leak into reports?
Test records are clearly labelled and archived or deleted after the check.