Glossary

On-call

On-call is a rotation that designates who is responsible for responding to incidents at any given time.

On-call is the practice of always having a designated responder reachable to handle incidents. A rotation shares this duty across a team over time — daily or weekly shifts — so no single person carries the pager indefinitely. Paired with an escalation policy, it guarantees that an alert always reaches someone able to act, at any hour.

Why it matters: detection is worthless if no one responds. On-call is the human half of incident response, and it is what keeps a 3 a.m. outage from waiting until morning. A well-run rotation also distributes the burden fairly, which protects the team from burnout and keeps response quality high over the long term.

A common misconception is that on-call is just a phone number to call. In practice it depends on an escalation policy — automatically paging a backup if the primary doesn't acknowledge within a set time — and on alert quality. On-call only works when alerts are fast and trustworthy; flood responders with false positives and they tune out, which lengthens acknowledgement time and defeats the entire purpose.

Sustainable on-call is as much an organizational design problem as a technical one: reasonable rotation length, clear handoffs, follow-the-sun coverage where possible, compensation or time-off for shifts, and ruthless reduction of noisy alerts so every page genuinely deserves a response.

FAQ

Frequently asked questions

What is an escalation policy?
A set of rules that decides who gets paged and in what order. If the primary responder doesn't acknowledge an alert within a defined time, it automatically escalates to a backup, ensuring nothing slips through.
How does alert quality affect on-call?
Heavily. Frequent false positives cause alert fatigue, so responders start ignoring pages and acknowledgement time rises. Keeping alerts accurate and actionable is essential for on-call to function.
How do you make on-call sustainable?
Keep rotations a reasonable length, share the load fairly, provide clear runbooks and handoffs, compensate for shifts, and continuously cut noisy alerts so every page is worth waking up for.

← All terms

Start monitoring in under a minute

Free forever, no credit card. Upgrade when you need finer intervals.