A well-designed HA cluster should have enough free resources to restart all of the business-critical virtual machines in the cluster, in the event of a host(s) failure. For this to be possible, it is essential for the cluster to maintain enough free CPU and memory resources, which are referred to as failover capacity. The failover capacity enables the restarting of the VMs while the cluster capacity is reduced, owing to the host failure.
A failover capacity that has been configured on a cluster determines the number of host failures that the cluster can sustain and still leave enough usable resources to support all the running and restarted virtual machines in the cluster. vCenter ensures that the configured failover capacity is maintained by using a mechanism called vCenter Admission Control.
We will cover more about this in the How it works...