The present solution provides a spillover management technique for virtual servers of an appliance based on health. Using a health based spillover technique, a network appliance may direct requests to a backup or second virtual server upon determining that a predetermined percentage of services being load balanced are down. In this manner, the spillover will occur based on a user controlled determination of a level of services being down to the number of services enabled. Instead of waiting for a last service of a virtual server to be marked down to spillover to another virtual server, the spillover may occur based on a user specified percentage. For example, the appliance may spillover from one virtual server to another virtual server when the number of services marked down relative to the number of enabled services falls below a specified percentage.