Skip to content

Maintenance Windows

Without a maintenance window, planned downtime looks exactly like an outage: your monitors go DOWN, alerts fire, and your status page shows red. Scheduling maintenance ahead of time keeps everyone informed without the noise.

Scheduling maintenance windows is available on the Max and Enterprise plans.

The Maintenance windows page
  1. Go to Alerting → Maintenance and click Schedule maintenance.
  2. Enter a title and an optional message for your users.
  3. Select the affected monitors.
  4. Set the start time and either a duration (30m, 1h, 2h, 4h, 8h) or a custom end time.
  5. Publish it.
TIMELINE MAINTENANCE 02:00 → 03:00 503 · 503 · 200 · 503200200 checks keep running · shown in blue · alerts suppressed · excluded from uptime % Before: announced on the status page as “scheduled”. After: auto-closes at the end time (or manually, earlier) — monitors return to their real status, alerts resume.
  • Affected monitors keep being checked as normal — monitoring never actually stops.
  • Their status is overridden to a dedicated maintenance state (shown in blue, distinct from both “up” and “down”) on the dashboard, the monitor list, and the public status page.
  • Alerts for those monitors are suppressed for the duration of the window — a check failing because you’re mid-deploy won’t page anyone.
  • The status page shows the maintenance window in advance if it’s scheduled for the future, and live while it’s in progress.

A maintenance window closes automatically once its scheduled end time passes — there’s nothing to remember to turn back off. If work finishes early, you can also end it manually from the maintenance list.

  • Schedule maintenance before you start deploying, not after alerts have already fired.
  • Keep the message specific (“Database migration, ~30 min of elevated latency”) so subscribers know what to expect.
  • If a maintenance runs long, extend it rather than letting it auto-close and alerts start firing again mid-work.