On-Call
On-call is a rotating arrangement in which engineers stand ready to respond to alerts and incidents outside normal hours, providing continuous coverage for reliable services.
On-call is the practice of having designated engineers ready to respond to alerts and incidents, including outside normal business hours. Responsibility usually rotates among team members on a published schedule so that coverage is continuous without burdening any one person.
How It Works
A rotation assigns a primary responder for each shift, often with a secondary as backup. When monitoring detects a problem that needs human attention, an alerting system pages the on-call engineer, who acknowledges the page and begins incident response using runbooks and dashboards. Effective on-call depends on alerts that are actionable and urgent; paging someone for issues that are neither erodes trust and causes fatigue. Healthy rotations cap the number of pages per shift, compensate on-call work, and feed recurring pages back into engineering work to remove their root cause.
Why It Matters
Services that users depend on at all hours need someone able to act when they fail. On-call provides that human safety net. Done well, it keeps services reliable and gives the engineers who build a system a direct stake in its operability, which improves the system over time. Done poorly, with noisy alerts and excessive load, it causes burnout and attrition. Managing on-call load, much of which is toil, is therefore a key responsibility for any operations-minded team.
Related Terms
On-call is the front line of incident management and a common source of toil. It contributes to lowering mean time to recovery and feeds postmortems.