How maintenance windows work

A maintenance window temporarily stops selected uptime checks and heartbeat failure alerts. Use one for planned deployments, restarts, backups, and other work that may cause expected downtime.

Maintenance windows repeat on a schedule. You choose the start time, timezone, duration, weekdays, checks, and heartbeats.

The recurring schedule

OnlineOrNot starts the window on the selected weekdays at the chosen local time. The timezone keeps that time consistent when daylight-saving time changes.

Start and end times may be about a minute late. OnlineOrNot only uses the time from the selected start date. The calendar date does not delay the first run, so the window can run earlier if a chosen weekday comes first.

What happens when a window starts

When the window starts, selected checks and heartbeats enter MAINTENANCE unless they are already paused or in maintenance.

Uptime checks stop running. Heartbeats stop becoming down when a ping is missed. This prevents new incidents and alerts for the planned work.

This does not turn on the check or heartbeat's manual Paused setting.

Start the window before maintenance begins. It will not close an incident that is already open, including an automatic status page incident. Manage that incident separately.

Heartbeat pings during maintenance

OnlineOrNot still records heartbeat pings during maintenance. A ping does not move the heartbeat out of maintenance or send a recovery alert.

You can still see that the job ran without receiving missed-ping alerts.

What happens when a window ends

When the window ends, checks and heartbeats return to PENDING:

• uptime checks start running again
• heartbeats wait for their next ping
• manually paused checks and heartbeats stay paused

This prevents the end of maintenance from sending a false recovery alert.

Maintenance, pause, and mute

Control	Intended use	What happens
Maintenance window	Planned work on a schedule	Selected checks and heartbeats enter maintenance, then return to pending
Pause	Stop monitoring until you resume it	Checks stop running and heartbeat pings are ignored
Mute	Keep a heartbeat reporting without missed-ping alerts	Pings remain recorded while missed-ping state changes stop

Use a maintenance window for planned work on a known schedule. Use pause when you do not know when monitoring should resume.

Scheduling boundaries

Use windows that start and end on the same local day. Do not schedule a window across midnight.

Do not overlap windows for the same check or heartbeat. One window may return it to pending while another window is still active.

Only edit, disable, or delete a window while it is inactive. Changing a running window can leave checks or heartbeats in the wrong state.

Monitoring maintenance vs. status page maintenance

Monitoring maintenance windows and status page scheduled maintenance are separate features:

• Maintenance windows stop monitoring for selected checks and heartbeats.
• Status page scheduled maintenance shows planned work to customers and can notify subscribers.

Creating one does not automatically create the other. Set up both when planned work should stop alerts and appear on a public status page.

To manage maintenance windows with the API, see the Maintenance Windows API reference.