# Uptime monitoring (/docs/databases/monitoring)

backupd checks every connected database every 5 minutes and alerts you when one becomes unreachable. Open a database and choose the **Monitoring** tab to see the results.

## What a check does [#what-a-check-does]

A check only opens a connection to the database's host and port and closes it again. It never logs in and never uses your credentials. For Convex, the check is an HTTPS request to the deployment. A check that gets no answer within 5 seconds counts as down.

Because a check only tests reachability, a database can be **Operational** while backups still fail, for example after a password change. Backup failures have their own alert.

## The Monitoring tab [#the-monitoring-tab]

- **Status:** **Operational**, **Unreachable** (with how long it has been down), or **Waiting for the first check**. The first check runs within 5 minutes of adding a database.
- **Check now:** runs a check straight away.
- **Uptime** over 24 hours, 7 days, 30 days and 90 days.
- **90-day history:** one cell per day.
- **Latency:** how long each connection took.
- **Incidents:** runs of consecutive failed checks over the last 30 days.

## Alerts [#alerts]

When a database goes from up to down, backupd sends the **Database unreachable** event to your notification channels. See [Notifications](/docs/notifications).