Designing a system with redundant parts and automatic failover so it stays up when a component fails, measured against an agreed uptime target.
High Availability (HA) architectures remove single points of failure. For AI that operations depend on, HA means redundant LLM inference nodes, load balancers, and Vector Databases distributed across multiple geographic zones, so the service keeps running through major hardware failures.
See also Air-Gapped Environment, Role-Based Access Control and Zero Trust Architecture.