Software Engineering · Distributed Systems
Automatically switching operation to a standby system the moment the primary one fails, with minimal disruption.
Mental model: Like a generator kicking in the instant the power goes out.
Example: A database's read traffic automatically shifting to a replica when the primary instance stops responding.
In practice: Automatic failover requires health checks and a way to redirect traffic — like a load balancer or DNS update — fast enough that clients barely notice the switch.