CLUSTER CREATION AND MANAGEMENT FOR WORKLOAD RECOVERY

Aspects of the disclosure relate to managing migration of one or more applications from a primary computing device to recovery computing devices using a controller. Resource data that includes application resource requirements and resource capacities is monitored. An application exists as a single instance on the primary computing device. A recovery cluster for the application consisting of recovery computing devices is determined. A division of the application into a plurality of application instances is identified. In response to the resource data being updated, a new recovery cluster is determined. In response to the new recovery cluster, a new plurality of application instances is identified. Once a triggering event on the primary computing device is detected, the controller migrates the new application instances to the new recovery cluster. Other uses of the disclosure are possible.