Chaos Engineering in the Cloud-Native Era: Evaluating Distributed AI Model Resilience on Kubernetes

In this study, participants assigned the role of a cloud-native system-operator will apply chaos engineering practices on Kubernetes clusters to assess the operation of distributed AI models' resilience.Chaos engineering is the most vital technique when it comes to testing the failure conditions of AI systems and their vulnerabilities.The study aims at analyzing how the concept of chaos engineering can be applied, specifically in Kubernetes environments and mainly in terms of fault tolerance and performance recovery.This study will explore how Kubernetes can enable the daily operations of AI applications at scale by performing multiple failure simulations and utilizing automated tools.It also outlines implements, techniques and approaches to use when performing chaos experiments in the context of this discussion.