|docs
Start Trial

Feature Tutorials

Maintenance Windows

Maintenance Windows Tutorial

View as Markdown

When a service goes down for planned work — a deploy, an upgrade, infrastructure maintenance — its monitoring keeps working exactly as designed, which means it keeps firing alerts. Without a way to signal "this downtime is expected," every planned maintenance window turns into a flood of false-positive incidents that eat into responders' attention and erode trust in alerts generally.

Maintenance Windows let you temporarily suspend a service's ability to trigger new incidents for a defined start and end time. While a service is in a maintenance window, its integrations are effectively switched off and no new incidents are created. Note that existing open incidents are not automatically resolved when a window begins — they stay open and continue notifying as usual unless resolved manually — and while a service is under maintenance, incidents on it can't be resolved through the Events API.

You can create, edit, and review maintenance windows from the Service Directory, which has a dedicated Maintenance Windows tab showing current, upcoming, and past windows, or from the PagerDuty mobile app. Maintenance windows are also fully scriptable through the REST API and Terraform, which is useful for wiring maintenance directly into a deploy pipeline so windows open and close automatically around a release.

Learn More

Read more about what Maintenance Windows provides

Find the product documentation in our Knowledge Base

Using the API?

Managing PagerDuty with Terraform?

Terraform resource: pagerduty_maintenance_window.

Find the full documentation for the PagerDuty Terraform Provider in the Terraform Registry