Maintenance Windows
Schedule maintenance windows to prevent false alerts during planned downtime and communicate with your users.
By defining maintenance windows in advance, you prevent false alerts during planned work and keep your team focused on real incidents.
Why Maintenance Windows Matter
Planned work looks like an outage to a monitor. A maintenance window tells Uptime the downtime is intentional, which does two things:
Keeps alerts for real incidents — your team isn't paged for work it scheduled, so an alert keeps meaning something is wrong
Tells your users ahead of time — your status page announces the work before it starts
Prevent False Alerts
When a monitor is in a maintenance window with alert suppression enabled, Uptime continues checking the service but won't send alerts. Your team stays focused on real problems instead of dismissing expected outages.
Communicate with Users
Maintenance windows can be displayed on your public status page, letting users know about planned work before it happens. This reduces support tickets and builds trust through transparency.
Keep a Complete Record
Checks keep running during a maintenance window, and their results count toward uptime like any other check, so your history shows exactly how the service behaved while the work happened. Reports list the maintenance minutes that covered each monitor beside its uptime, so you can tell planned downtime apart when you share the figures.
Configuration Options
When creating a maintenance window, you can configure the following options.
Name
A descriptive name for the maintenance window, such as "Database migration" or "Weekly infrastructure updates". This name appears on status pages, in their maintenance calendar, and in your list of maintenance windows.
Scope
Choose how wide the maintenance window reaches:
All monitors — Every monitor you have, of every kind
A kind of monitor — Every HTTP, DNS, TCP or flow monitor, or every cron job, including ones you create later
A single monitor — One monitor, picked from the monitors of its kind
Narrow the scope when only some services are affected. If you're upgrading a database server, apply the window to the monitor that depends on that database.
A whole kind is the right choice when the work happens below your services rather than inside one of them. Migrating a zone to a new nameserver puts every DNS monitor in doubt at once, and a DNS-wide window quiets all of them — including the records you add mid-migration — without touching the HTTP checks that should still page you if the site itself goes down.
Start and End Time
Starts at and Ends at set when the window begins and ends, and the end must come after the start. Both are read in UTC, whatever your computer's time zone. You can schedule windows days or weeks before the actual work.
Choose an end time that covers your expected maintenance plus some buffer. You can always end early if you finish ahead of schedule.
Suppress Alerts
When enabled, no alerts are sent for monitors covered by this maintenance window. Uptime still performs checks and records results, but your team won't be notified.
Keep this enabled for routine maintenance. Disable it only if you want to monitor the maintenance process itself and get alerts if something goes unexpectedly wrong.
Show on Status Page
When enabled, the maintenance window appears on your public status page. Users see:
Upcoming maintenance with the scheduled time
Active maintenance while work is in progress
The maintenance name and affected services
An "Add to calendar" link that downloads the window as an
.icsfile
Readers can also subscribe to the status page's maintenance calendar, which lists every window shown there and updates in their calendar app when you move, end or delete one.
Enable this for any user-facing maintenance. Disable it for internal work that doesn't affect user experience.
Recurring Maintenance
For regular maintenance schedules, turn on Recurring and write a pattern in Recurrence pattern (when recurring):
Daily:
daily:02:00-04:00Weekly, on one day:
weekly:monday:02:00-04:00Monthly, on one date:
monthly:15:02:00-04:00
Recurring times are in UTC. An end time earlier than the start time runs past midnight, so 23:00 to 01:00 is a two-hour window. A monthly window on the 29th, 30th or 31st falls on the last day of shorter months.
Recurring windows are useful for:
Weekly deployment windows
Nightly backup windows
Monthly patching schedules
A recurring window is one entry in your list: editing or deleting it changes every occurrence.
Maintenance States
Maintenance windows move through three states.
Upcoming
The maintenance window is scheduled but hasn't started yet. During this state:
The window appears on status pages (if configured)
Monitors operate normally and send alerts
You can edit or delete the window
Active
The maintenance window is currently in progress. During this state:
Alerts are suppressed (if configured)
Status pages show maintenance in progress
Uptime continues monitoring and recording check results
You can end the window early
Past
The maintenance window has ended. Past windows:
Are kept in your maintenance history
Can be reviewed for auditing purposes
Show which checks occurred during the window
Ending Maintenance Early
If you complete maintenance ahead of schedule, open the window with Edit and click End maintenance now to:
Resume normal alerting immediately
Update status pages to show services are operational
Accurately reflect the actual maintenance duration
Don't leave maintenance windows running longer than necessary. Ending early ensures you're alerted to any real issues that might occur after maintenance completes.
Best Practices
Always Schedule Before Work Begins
Create maintenance windows before starting planned work, not during. This ensures:
Status pages are updated before users notice downtime
Your team knows not to respond to expected alerts
Maintenance is documented in your history
Enable Status Page Visibility
For any maintenance that affects users, enable Show on status page. Proactive communication:
Reduces support ticket volume
Builds user trust
Demonstrates operational maturity
Use the End Early Feature
When maintenance completes, end the window immediately. This:
Restores normal monitoring and alerting
Shows users that services are back online
Keeps your maintenance records accurate
Set Up Recurring Windows for Regular Maintenance
If you have scheduled maintenance periods (weekly deployments, nightly backups), create recurring windows. This:
Reduces manual work
Ensures maintenance is never forgotten
Provides consistent communication to users
Apply to Specific Monitors When Possible
If only some services are affected, select specific monitors rather than applying to all:
Other services remain fully monitored
Status pages accurately reflect what's affected
Next Steps
Understanding Monitors — Learn about the different monitor types
Understanding Alerts — Configure how you get notified
Status Pages — Communicate service status to your users