# Incident (IT)

> A problem no one planned. It breaks a service or makes it worse for users.

**Field:** Infrastructure  
**Category:** Reliability  
**Also known as:** Outage, IT Incident, Production Incident, Incident Response

## Think of it like

It is like a fire alarm going off. Everyone stops, and one person takes charge.

## Definition

An incident is an unplanned event that breaks a service or makes it worse for users. It could be an outage, very slow pages or lost data. Teams rank incidents by how bad they are, often as sev 1, sev 2 and so on. A big one gets an incident lead who runs the response. The first goal is to get users working again. Finding the root cause comes after.

## Example

At 2 p.m. checkout starts failing for every buyer. Jack calls a sev 1 incident, rolls back the morning release, and checkout works again by 2:20.

## Why it matters

A clear plan for incidents turns panic into routine. Users get back to work sooner.

## Common mistake

Hunting for the root cause while users are still down. Restore the service first, and dig in after.

## Three principles

- **Users first.** Restore the service before finding why.
- **One lead.** A single person runs the response.
- **Ranked by harm.** Severity levels set how fast to act.

## Related terms

- [On-Call](https://learn.benibix.com/term/infrastructure/on-call)
- [Runbook](https://learn.benibix.com/term/infrastructure/runbook)
- [Postmortem](https://learn.benibix.com/term/infrastructure/postmortem)
- [Mean Time to Recovery (MTTR)](https://learn.benibix.com/term/infrastructure/mean-time-to-recovery)
- [Alerting](https://learn.benibix.com/term/infrastructure/alerting)
- [Rollback](https://learn.benibix.com/term/development/rollback)
- [Incident Commander](https://learn.benibix.com/term/infrastructure/incident-commander)

## Told apart from

- [AI Incident vs Incident (IT)](https://learn.benibix.com/compare/ai-incident-vs-it-incident): People think an AI incident is an IT outage, and it is not. An IT incident is a service that breaks or slows down. An AI incident is harm caused by what an AI system did, such as a wrong answer someone acted on. Every server can be up while the harm happens. So the usual uptime alarms will not catch it.

![A timeline from left to right. Something breaks. An alert fires and pages the person on call. They open the runbook and fix it, and the service is back. The span from break to back is marked as time to recover. After that, a postmortem is written, and its actions feed back to the start.](https://learn.benibix.com/diagrams/inf-incident-timeline.svg)

---

Source: https://learn.benibix.com/term/infrastructure/incident
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-19
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
