Incident management
Incident management is the process an engineering team follows when a service is disrupted or degraded. It covers detecting the problem, declaring an incident, coordinating the response, restoring service and learning afterwards. The goal is to limit user impact and recover quickly, with clear roles instead of everyone talking at once.
How it works
A typical lifecycle has five steps:
Detect: an alert or a user report shows something is wrong.
Declare: someone opens an incident and sets a severity, for example SEV1 for a full checkout outage.
Respond: an incident commander coordinates, an operations lead works the fix, and a communications lead posts updates, say every 30 minutes.
Mitigate and resolve: roll back, fail over or scale first, then fix the root cause.
Review: write a blameless postmortem with owned action items.
Why it matters in production
Without a process, the loudest person drives the response and fixes collide. Clear roles and severity levels cut mean time to recovery and keep stakeholders informed. The review step turns each outage into fewer future ones.
See also
On the One2N blog
Further reading
