An important consideration for cluster design is planning for server failure, or planned maintenance. vSphere clusters are typically used to provide increased availability and resource load balancing, and are a key component of a VMware Cloud Provider Program platform design. One part of the cluster design, discussed previously, is the proportion of the cluster’s total capacity that is reserved for a failure event or reserved to undertake operational maintenance tasks. Determining the number of compute nodes in each cluster and the total capacity you want to reserve (which makes it unavailable for tenant workloads) will provide the percentage of capacity that must be allocated for failure scenarios.