Live Experiments with K8s Applications - Pitfalls and How to Avoid Them

Explore the intricacies of live experiments with Kubernetes applications in this informative conference talk. Discover common pitfalls and effective strategies to avoid them when implementing metrics-driven automated rollouts using canary, A/B, and A/B/n comparisons. Learn how subtle differences in experiment design, including metric collection, traffic shifting policies, and experiment duration, can significantly impact outcomes and production decisions. Gain insights into statistically effective remedies that enhance experiment repeatability and accuracy, ultimately leading to more trustworthy results. Delve into key topics such as current trends, principles of trust, and essential terminology. Follow along with a comprehensive demo that illustrates these concepts in practice, and walk away with valuable knowledge to elevate your CI/CD processes in Kubernetes environments.