Glossary · Incidents and alerting

What is a maintenance window?

A maintenance window is a scheduled period for planned work, such as upgrades, migrations or deploys, during which a service may be unavailable and alerts are paused on purpose.

How it works and why it matters

Windows are announced in advance, usually on a status page and by email, and timed for low traffic. Many SLAs leave announced maintenance out of downtime, often with limits on how long and how often.

Pausing alerts during a window keeps the team’s attention on the work instead of on alerts they already expect. Pausing them only for the affected services, and only for the planned time, keeps real problems elsewhere visible.

Work that runs past its window should be treated as an incident from that moment, and the status page updated.

How Uptimecheck helps

Monitors in an Uptimecheck maintenance window aren’t checked, open no incidents, send no alerts and lose no uptime. Status pages show upcoming and ongoing maintenance and can email subscribers; a window can repeat daily, weekly or monthly with a reminder before each one, and you can start one from a deploy script through the API.

See Incident management

14-day free trial · No credit card required

Read more in the guides

Questions

Does a maintenance window count as downtime?

That depends on your SLA’s definition; many leave announced windows out. In Uptimecheck, time in a maintenance window doesn’t count against uptime.

How far ahead should maintenance be announced?

At least a few days for work customers will notice, and longer for big changes. Send a reminder shortly before it starts.

What if maintenance runs late?

Extend the window, say so on the status page, and treat any part that runs past the announced end as an incident.

Downtime is obsolete.

Set up your first monitor in under a minute, or look around the live demo first.

14-day free trial · No credit card required