Software Performance testing is one of the important aspects of testing a product. It is more important when we have a cloud application, i.e. not a standalone application but a web application. Many of us use various tools like JMeter to test the performance of different web requests. But how do we analyze the results of JMeter? A failed Request is easy to catch. What about performance degradation? How do we know if a request performed worse than the previous runs or not? How do we take into consideration other environmental parameters – memory, processing power, operating system limits, etc? Performance benchmark using machine learning is one of the ways using which we can take care of some of these manual study and detection of performance regressions which may otherwise go unnoticed.