Early identification of anomalies from user experience data

Detecting anomalies in time series data is a critical task in areas such as cloud health monitoring. This Thesis proposes a proof of concept for forecasting and detecting anomalies in time series data. The proposed approach is based on Facebook Prophet model which is an open source library built on decomposable (trend+seasonality+holidays) models. It gives the user the power to perform time series predictions using simple intuitive parameters with acceptable prediction result. Moreover, the architecture helps the concerned team to detect outliers and understand what kind of problems that they may have. The results on Cloud Functional Testing data show the ability of the proposed model to detect anomalous patterns in time series from different fields of application. This study presents the capability of accurately forecasting future cloud health with expected level of reliability in our forecast