Abstract:
"
In recent years, microservices have been more familiar with software developers
because it is simple to code and get the job you need while comparing to monolith
applications. As a result, CI/CD comes into play to ensure continuous off-app delivery
when companies follow agile developments. This CI/CD environment consists of
several phases, such as error code testing, integration testing etc. They do not have a
simple way to perform performance testing and get the basic idea of the overall
system throughput.
This research is focused on developing a microservices performance degradation
detection system using machine learning and created a simple and straightforward
mechanism. Based on the existing research author decided to try out supervised
learning methods using XGBoost. The final solution was able to detect performance
degradation. It helps DevOps and cloud engineers determine how they should deploy
the system (whether increase the virtual machine's performance or decrease it) and
adjust auto-scaling values based on the result predicted.
Keywords— CI/CD, Performance Degradation detection, microservice, metric"
