Links
Uptime monitoring
One URL per app, checked every minute, alerting exactly once when it goes down and once when it recovers.
Dashboard → Issues → Uptime. Point it at the endpoint whose failure means your app is down — usually your API's health check.
- Cadence — a check every 60 seconds with a 5-second timeout.
- Down means three consecutive failures — a single blip never pages anyone.
- Alerts fire exactly once per transition —
uptime.downwhen it trips,uptime.recoveredwhen it clears — through Slack, email, and your webhook. - History — checks are retained 30 days, and the Issues page shows 7-day uptime next to your error feed, because the two questions are usually asked together.
Monitor the dependency that costs you money
If your app's core loop rides a vendor API, monitoring your own endpoint that exercises it catches the outage your users actually feel — not just whether your server process is alive.