Incidents · Starter+
Scheduled maintenance
Schedule the window today, sleep through the 3 AM migration, wake up to a green status page that already explained itself.
Why you need it
Surprise downtime is an avoidable support fire. A scheduled-maintenance incident pre-announces the window, opens itself at the start time, and resolves itself when the work finishes — so the customer-facing communication happens whether you're awake for it or not.
- Upcoming banner customers see ahead of time. Windows within seven days surface on the public page above the service list. Customers plan around them; you stop fielding "is the migration tonight still happening?" tickets.
- Auto-open, auto-resolve. The window opens itself at start, flips affected components to their impact level, and closes itself at end with a "maintenance completed" update. No human alarm clock required.
- Maintenance subscribers are a separate opt-in. Customers who want incident alerts but not every Sunday cleanup notification can opt out of maintenance specifically — your unsubscribe rate stays low.
- SLA-aware by default. Planned windows skip the public SLA calculation, so a quarterly migration doesn't quietly drag your 99.9% number down.
- Manual override always available. Resolve early when the work finishes ahead of schedule, extend End before it lapses if the rollback takes longer — the activator never overwrites your decisions.
Where it pays off
The right tool any time customers need to know about downtime before it happens:
- Overnight database migration. Schedule the window today, list affected components, customers see the banner 72 hours out, you sleep through the auto-open and auto-resolve.
- Recurring cleanup windows. Sunday 02:00 UTC every week. Open a fresh incident from a template; the automation handles the rest.
- Vendor-imposed maintenance. Your hosting provider announces a region restart. Mirror the window on your page so customers see one consistent story instead of two.
- Pre-deploy quiet zones. A 30-minute Maintenance window during a release prevents the auto-incident engine from firing on a probe flap during the deploy.
Not the right tool when: the work is genuinely unplanned — open a regular incident in Investigating instead of backdating a "maintenance window" no customer ever heard about.
Available on Starter+. Already on StatusPulse? See the full config in Help →
Related
Try Scheduled maintenance in StatusPulse
5 probes, 1 status page, forever. No credit card. US or EU host — you choose.