Traditional testing strategies often fall short in preparing systems for the unpredictability of real-world cloud environments. In this talk, Shivendu shares how his team built a Kubernetes-native chaos engineering framework to proactively test and improve system resilience. By leveraging a custom Kubernetes operator, cluster manager, database pods, and load generators, they created a robust platform to simulate failure scenarios and observe system behavior under stress. The session details the implementation of automated chaos experiments, real-time health checks, and observability using Grafana, Loki, and Prometheus. Shivendu also reflects on the broader organizational impact, including key lessons learned, challenges faced, and how chaos engineering practices helped strengthen both the product and team processes. Attendees will gain practical insights for designing their own experiments to build more resilient, cloud-native systems.