Fault tolerance is a paradigm for providing high availability in a computerized system where application services are replicated to multiple nodes. The paradigm is widely used in a cloud computing environment where users may benefit from automatic backup when an application such as a clinic support system is deployed in the cloud. However, applications residing in premises such as small clinics are still prone to an outage and would require certain time with human involvement when performing recovery. Thus, this research aims to provide a prototype of a backup mechanism for a health system residing in-premise of a clinic. The study will also act as a feasibility investigation of fault tolerance mechanism in edge of network where we proposed the use of fog computing model that acts as middleware for detecting and failover solution when an outage has temporarily occurred. The middleware will perform failure detection through heartbeat and replicate the services at the same time. When an outage is detected, the middleware will take over as a secondary service provider to ensure applications may be used seamlessly. There are four test simulations for testing the proposed mechanism where three has shown successful recovery actions and only one fail test case due to limitation at the client side.