From Monolith to Microservices : building scalable applications with Kubernetes, CI/CD, and chaos engineering

Abstract

This thesis explores the development and resilience evaluation of an application designed using a microservices architecture and deployed on a local Kubernetes cluster. The primary objective was to assess the application's ability to withstand various failure scenarios by applying chaos engineering principles, using the Chaos Mesh platform. The research involved the systematic injection of faults – including pod failures, network disruptions and resource constraints – to uncover vulnerabilities and evaluate the system’s fault tolerance.

The experiments revealed critical issues, such as concurrent user account creation failures under network latency and exposed the differing resilience level of different technology stacks within the application. Additionally, a questionnaire was conducted to assess awareness and perceptions of chaos engineering within the IT community. The responses indicated growing interest, despite limited widespread familiarity.

The findings of this work demonstrate the practical value of chaos engineering in uncovering systemic weaknesses and underscore the importance of sound, proactive resilience strategies in modern microservices-based systems. This work contributes to a deeper understanding of how to build and test resilient cloud-native applications and highlights the potential for broader adoption of chaos engineering practices within the IT industry. It also provides valuable insights for universities and technical education institutions that may wish to incorporate resilience engineering concepts into their software engineering and DevOps curricula.