When a service fails
If a website goes down at night, an alert goes to the person on call. They check who is affected, inform colleagues, and work to restore the service.
A runbook lists steps for a known problem: how to check it, limit harm, restore service, and confirm the fix. Escalation means bringing in someone with the skills or authority needed to help.
Share the work
Teams limit alerts to urgent problems, give on-call staff training and access, and share shifts fairly. After an incident, they use what they learned to improve the service and runbooks.