Skip to content
Last updated July 2, 2026

Maintenance windows

Maintenance windows tell HostAtlas “expect noise here — don’t page anyone.” Monitoring keeps running; alerts and notifications are suppressed for the scope you choose, then automatically resume when the window closes.

Under Alerts → Maintenance (or /maintenance) you get a table of windows:

  • Title and optional description.
  • Scope: a single server, a group of servers by tag, or the entire fleet.
  • Start / end times (UTC and local).
  • Status badge: scheduled, active, or ended.
  • On / off toggle.

Active windows are highlighted; the servers they cover show a small “under maintenance” chip on the Servers list and on the server detail page.

  • New window — pick scope, start, end, title. Optional: a note for the audit log.
  • Edit — reschedule or change scope on an existing window.
  • Toggle — flip a window off without deleting it (useful for one that’s on the calendar but no longer needed).
  • Delete — removes the window; any suppression it was providing ends immediately.
  • While a window is active, every alert rule that would target a server inside its scope is silenced. The rule still evaluates, an event may still be recorded internally for context, but no notification goes out and no incident is opened.
  • Suppression starts and stops automatically at the window’s timestamps — you do not need to press anything at start or end.
  • Maintenance status is visible everywhere the affected servers appear, so operators can see at a glance why a normally-loud host is quiet.
  • Scope options:
    • Server — one host.
    • Tag — every server carrying a given tag.
    • Global — every server in the workspace (use sparingly).
  • Alerts & Incidents — the rules whose notifications windows suppress.
  • Recovery rules — combine with windows so auto-actions do not run during planned work.
Was this page helpful?