Learning about an outage from a customer is the worst case.
Free monitoring
- UptimeRobot — check every 5 min
- StatusCake
- Better Stack
What to track
- HTTP status (200 or 503)
- Page load time
- SSL expiry
- Domain expiry
Alerts
Email, SMS, Telegram, Slack — use multiple channels.
Status page
Big services have status.example.com — customers see real status.