Single point of failure
A single point of failure, also known as SPOF, is any component of a system that causes the whole system to stop working if it fails. When designing reliable systems, SPOFs can be avoided by implementing redundant components and replicating critical parts of the system. For example, a computer may implement RAID storage so that if a single disk fails, the system can continue operating.
On a larger scale, a website may replicate application servers. If a single server encounters a critical error, service automatically switches to another server, and end users are not prevented from using the application.