Guides

On-call and escalation basics

Last updated 23 June 2026

How on-call rotations, escalation ladders and acknowledgements fit together, so the right person is reached when something breaks, without burning out the team.

On-call is how a team makes sure someone is responsible for responding when something breaks, around the clock if needed. Escalation is what happens when that someone does not answer. Done well, the right person is reached quickly and everyone else is left alone. Done badly, alerts go to a group chat that no one owns, and real problems sit until a customer complains.

Rotations

A rotation shares the on-call duty across a team on a schedule, so no one person carries the pager forever. A simple weekly handoff is enough to start. Larger teams layer rotations (a primary and a secondary, or a follow-the-sun setup across time zones) so coverage matches when people are awake.

Escalation policies

An escalation policy is the ladder an alert climbs until someone acknowledges it. A typical one: page the on-call person; if they do not acknowledge within a few minutes, page them again or page the secondary; if still nothing, page a wider group or a manager. Each step has a wait, so a missed phone gets backstopped automatically instead of an incident going unanswered.

Acknowledge to stop the noise

Acknowledging an alert tells the system a human has it, which stops the escalation from climbing further. This is the mechanism that lets you be aggressive about paging without spamming the whole team: the ladder only advances while no one has taken ownership.

Severity and channels

Not every alert deserves a 3am phone call. Route by severity: a major outage can ring a phone and send an SMS, while a minor degradation might only post to a channel or email. Matching urgency to impact is the difference between a pager people trust and one they mute.

Coverage and humane defaults

Plan for real life. People go on holiday, so let someone pause their own paging while the rest of the rotation still covers the gap. Watch for coverage holes where a schedule resolves to nobody, and make sure an unacknowledged page eventually reaches a human rather than falling into silence. Alert fatigue is a real failure mode: every page that was not worth waking up for makes the next real one easier to ignore.

Set up on-call rotations and escalation ladders that reach the right person.

Start free

Frequently asked questions

What is an escalation policy?
The ordered ladder an alert follows until someone acknowledges it: page the on-call person, wait, then page again or page a secondary, then widen to a group. Each step has a timeout so a missed alert is backstopped automatically.
What does acknowledging an incident do?
It tells the system a human has taken ownership, which stops the escalation ladder from climbing further. It is what lets you page aggressively without spamming the whole team.
How do on-call rotations work?
A rotation shares on-call duty across a team on a schedule (often a weekly handoff) so no one carries the pager indefinitely. Larger teams stack a primary and secondary, or run follow-the-sun across time zones.
How do I avoid alert fatigue?
Route by severity so only real, high-impact problems wake someone, confirm failures before paging so blips never do, and keep rotations and holiday cover honest so the load is shared.