Proactive Fault Prediction in Microservices Applications Using Trace Logs and Monitoring Metrics

An increasing number of developers and researchers are embracing microservices due to the scalability and flexibility benefits they provide. Fault prediction in microservices architecture based systems is crucial for minimising computation costs and reducing development time. Early prediction of faults and their locations during the system development life cycle simplifies maintenance and enhances resource utilisation. To our knowledge, little to no effort has been made in predicting faults beforehand, instead most of the research that has been done thus far has been on detecting faults after they occur. Early error detection allows for proactive remediation, avoiding any negative impact on the performance of the application. This paper fills this gap by proposing a fault prediction model based on trace logs and monitoring metrics for microservices applications. The paper has inspected two different models for fault prediction of microservices application and the forecasting accuracy of these models is compared by using standard performance measures. A case study microservices application is used for empirical analysis of these models and the results indicate that RNN predicts the faults with minimum errors and higher accuracy.