Static analysis techniques and tools for infrastructure as code on the Kubernetes platform

This thesis investigates Infrastructure as Code, mainly in the Kubernetes platform. More specifically, it starts by explaining Infrastructure as Code and then deep dives in the Kubernetes architecture and the technologies around it, like Docker, Helm Charts, static analysis tools like Checkov, Terrascan and more. Technologies mentioned above are the introduction for the main focus of this thesis, which is static analysis and refactoring of the Kubernetes manifests for maintainability purposes. In that scope tools like Ansible, Terraform, Pulumi and others are analyzed as well. Building upon this knowledge this study examines the challenges related to Kubernetes manifests, like issues related to security and repetitive code blocks among others. It reviews existing static analysis approaches and tools aimed at improving code reliability, security and maintainability. The main contribution of this work is the proposal of a static analysis method. More specifically, this method suggests a way to refactor helm charts created by Helmify, which is an automation tool aiming to transform traditional manifests files to helm charts. The core target of this method is to present a way to reduce repetitive code as it occurs by the output of helmify. The core idea is to create a template, which will contain the union of all the features contained in the manifest files. Manifest files then have to be reconfigured to just reference this newly created template instead of containing all the feature they previously had. Overall the method proposed in this Thesis has been evaluated on some real world applications with results showing dramatic decrease in repetitive code, helping in the overall maintainability of the Infrastructure as Code.