Machine Learning applied to Kubernetes security in AWS

Abstract

The purpose of this work is to create a way to detect new attacks in Kubernetes by implementing an anomaly detection and alerting framework that targets the logs generated by the Kubernetes apiserver. The framework will use a machine learning anomaly detection algorithm and be implemented using AWS infrastructure. During the project we design and implement the anomaly detector and alerting logic and deploy it using the infrastructure that the system needs, which is also designed during the project.