Skip to content

Follow the Sun on-call

Follow the Sun is a schedule type you pick when creating a roster, not a separate product or a set of routing rules. The roster keeps its normal rotation - the same members, the same hand-off cadence - and you add working shifts for each region. When an alert fires while the on-call member is outside their working hours, the shift that is awake gets a first look on Slack. If nobody there takes it within the wait you set, the on-call member is paged exactly as the escalation policy says.

Regional working shifts hand alerts to whichever team is awake before paging the rota engineer

On-Call, then Rosters, then Add Roster, and choose Follow the Sun as the schedule type. The schedule type is fixed once the roster is created, so decide up front; converting a plain roster means creating a new one.

Fill in the rotation as you would for any roster (members, order, hand-off day and time, shift length - see Roster management), then add one working shift per region. Each shift has:

FieldMeaning
Name and timezoneFor example EMEA in Europe/London. Hours are interpreted in this timezone.
Working hoursStart and end times in the shift’s own timezone. An end time at or before the start means an overnight shift.
DaysWeekdays the shift works. Leave empty for every day.
MembersRoster members who belong to this shift.
Slack channelWhere nudges for this shift are posted. Optional; falls back to the roster’s default channel.

Roster-level settings control what happens around the shifts:

SettingDefaultMeaning
Wait for acknowledgement5 minutesHow long the awake shift has before the on-call member is paged. 1 to 60 minutes.
Page immediately for critical alertsoffSkip the nudge entirely for critical-severity alerts.
Gap-hours nudgepage immediatelyWhat to do when no shift is in working hours: page the on-call member at once, or nudge the roster’s default channel with a shorter wait (default 2 minutes) and a message saying nobody is on shift.
  1. The escalation policy reaches a level that targets the roster.
  2. AlertKick works out who is on call from the rotation and overrides, as for any roster.
  3. If that person is inside one of their working shifts, they are paged immediately. Nothing changes from a plain roster.
  4. If they are outside their shift, the currently awake shift’s Slack channel receives the alert with an Acknowledge button and the wait starts.
  5. Anyone acknowledging - in Slack, the dashboard, the mobile app, or over MCP - cancels the page. If the wait expires unacknowledged, the on-call member is paged on the channels the escalation level names, and the policy continues climbing from there.
  6. If no shift is awake (gap hours), the gap-hours setting decides between paging at once or a shorter nudge to the default channel.

Critical alerts can bypass the nudge with Page immediately for critical alerts, so a real outage never waits five minutes for a Slack reaction.

The roster timeline shows each shift’s working window as a tinted band behind the rotation bars, and a 24-hour coverage strip along the bottom shows handovers and gaps in your local time. Hatched stretches are gap hours: nobody on shift. If the strip shows a gap you did not expect, check for an overnight shift whose end time is on the wrong side of midnight, or a shift whose days list excludes the weekend.

  • A connected Slack workspace, because the nudge is a Slack message. See Notification channels for connecting one. Without Slack the roster still works; alerts skip the nudge and page directly.
  • Members with the right timezone on their profile, so quiet hours and the coverage strip line up with reality.

If you would rather stay on plain rosters, the older pattern still works: one roster per region with non-overlapping shifts, and escalation policy levels that fall through.

RegionHours covered (UTC)Roster
APAC22:00-06:00apac-oncall
EMEA06:00-14:00emea-oncall
AMER14:00-22:00amer-oncall

The escalation policy points at all three; only the active one has a member when the policy is evaluated. The trade-off is three rotations to maintain instead of one, and no Slack-first step.