When dealing with the reliability of a network, it is important to maintain the network components physically connected. When all the components are connected but then also the end user is not getting the service, this means that there is performance degradation. Such performance failure of a network occurs due to the imbalance of QoS parameters such as delay, jitters, throughput, etc. It is important to detect such failure and find the counter measures for it. A service-level agreement (SLA) is an agreement between two or more parties, where one is the customer and the others are service providers. Every service provider provides certain service contract or SLA to the customer that defines all the aspects of the service in terms of scope, quality, and responsibilities. The deterioration of SLA can also lead to performance failure. In a network, all the services should offer QoS guaranties that are resilient to failure too. The QoS connection requires that the existing routing mechanism should compute a path that satisfies QoS constraints. If all nodes and links are working fine but the performance of a network degrades, then too it is a network failure. Such type of failure is called as performance failure. Performance failure can be caused due to congestion, insufficient bandwidth, large delay in packet transmission, etc. These factors must be taken into account while detecting the network failure. Our notion is to detect such performance failure and provide a path that satisfies the QoS constraints without hampering the service provided to the end user. © 2015 IEEE.