How On-Call Scheduling Works

On-call schedules are attached to a team and define which members are responsible for responding to incidents during specific time windows. When an escalation policy step targets on_call, MonitoringDog resolves this to whoever is in an active schedule window at that moment.

Schedule Types

Date Range

A one-time window with explicit start and end dates. Use this for ad-hoc coverage — holidays, planned maintenance, or filling a gap in the rotation.

Recurring

A repeating window defined by an rrule pattern (e.g. "every Monday through Friday, 09:00–17:00, Europe/Berlin"). The rrule is expanded at query time to materialise the concrete on-call windows within any requested date range.

How "On-Call" Resolves

When an escalation step with contacts: on_call fires:

1. MonitoringDog checks for an active schedule window at the current timestamp.
2. The users assigned to that window are notified via the step's severity channels.
3. If multiple users are in the window, all of them are notified simultaneously.

What Happens if Nobody Acknowledges

If no one on-call responds, the escalation policy handles it automatically:

• Later steps with higher-severity channels (SMS, phone call) fire per their configured stepDelay.
• If repeatCount > 0, the entire policy sequence repeats.

On-call scheduling controls who gets notified. The escalation policy controls what happens when they don't respond.