Failover is switching a service to a backup system when the primary one fails. The switch can be automatic, triggered by health checks that detect the failure, or manual, where a person decides and carries it out. The backup may be in another availability zone, another region or another provider.

It matters because it determines how long an outage lasts. A tested failover can keep a service running or restore it quickly, which supports commitments such as availability targets in a service-level agreement.

Having a backup does not mean failover will work. The backup needs current data, enough capacity and correct configuration, and the switch itself can fail or take longer than expected. Automatic failover can also trigger when it should not, and failing over across providers adds extra complexity and cost, including data transfer between them. A plan that has never been tested is an assumption, not a guarantee.

To make it dependable, decide what level of downtime and data loss is acceptable, test the process regularly under realistic conditions and document who does what when it happens.