Building a self-service learning platform using cloud native tools and gitops

Abstract

Providing hands-onlaboratory exercises for developers is an effective way to build competence with newtools and methodologies. While many commercial and open source platforms offer learning paths for users, they often come at a cost and offer limited customization. Additionally, emerging methodologies such as chaos engineering often lack guided learning tracks altogether.

To address these gaps, this study explores the potential of using cloud native tools and methodologies in building a scalable learning platform in cloud. This study highlights how self-service learning platforms can streamline ordering of laboratory environments, reduce manual intervention by leveraging automated deployments, and automatically scale under growing demand.

This thesis is grounded in a comprehensive review of existing literature on current trends in cloud development, including the rise of cloud native practices and the adoption of open source technologies in the industry. Based on this research, a scalable learning platform with self-service capabilities is developed using curated cloud native tools. A novel chaos engineering lab exercise—featuring a reference application for fault injection and chaos engineering tools absent from other learning platforms—is designed to evaluate the built platform.

The learning platform maintained stability with multiple concurrent chaos engineering laboratories running chaos experiments simultaneously. The blast radius of individual fault injections was effectively minimized through strict access controls assigned to each user, ensuring that the disruptions remained contained within their respective laboratory environments. Even under heavy load with a continuously growing number of concurrent laboratory environments, the platform auto-scaled effectively to meet the increasing demands of user activity.

In summary, this thesis provides insights into the cloud native ecosystem and demonstrates how cloud native tooling can be leveraged to build automated, self service learning platforms. The implemented platform serves as a novel example of modern platform development using open source technologies, showcasing demand-driven auto-scaling for cost-efficient resource utilization and self-service capabilities with a high degree of automation to minimize manual intervention throughout the entire laboratory exercise.