Updated 2026-09-29
Planned work — a deploy, a database migration, infrastructure maintenance — shouldn't set off the same alerts as a real outage. A maintenance window tells Status about the work ahead of time.
Creating a window
A window has a name, a start and end time, and a set of checks it applies to — you can pin it to specific checks or apply it to your whole account. Optionally, a window can recur daily, weekly (on chosen weekdays) or monthly (on chosen days of the month), every N days, weeks or months, until an end date. Times are set in a timezone you choose (your account's by default).
What a window does
- Holds alerts by default for its checks while it runs. A check can still go down during the window (this is recorded in its history), but no alert goes out. If the check is still down when the window ends, the alert is sent then, so an outage that outlasts the work isn't missed.
- Shows on your status page by default, so visitors see a scheduled maintenance banner instead of an unexplained outage. Subscribers to that status page are notified twice: once when the window is scheduled (if it starts in the future), and again when it starts. A recurring window sends a fresh "starting" notice at each occurrence.
Each window has three switches, all on by default: hold alerts, show on the status page, and notify status-page subscribers. Turn any of them off per window.
When a window ends
Monitoring and alerting resume automatically the moment the window's end time passes — there's nothing to manually re-enable. If a check is still down at that point, its alert is sent then.
How many windows you can create
Your plan sets a limit on the number of maintenance windows you can have. A window that has finished no longer counts toward your plan's limit. See Choosing a plan for specifics.