CloudStack provides high availability (HA) by automatically restarting failed virtual machines on another healthy host within the same cluster. This is achieved through continuous host monitoring and VM heartbeat checks. If a host becomes unreachable, CloudStack triggers VM restart on another available node, reducing downtime and improving service reliability in production environments.