Zonal Outage Operational Stories in Kubernetes Clusters

Explore real-world zonal outage scenarios and their impact on Kubernetes clusters in this insightful conference talk by Jyoti Ranjan Mahapatra and Shyam Jeedigunta from Amazon Web Services. Gain valuable insights into the concept of availability zones as failure domains and learn how Kubernetes cluster administrators deploy components to achieve high availability. Discover the behavior of Kubernetes components during various types of zonal failures, from partial to full outages, including network partitions, power loss, reboots, and software deployment issues. Benefit from the speakers' extensive experience in operating a large fleet of Kubernetes control planes in AWS as they share operational stories and improvements that have enhanced resiliency for thousands of clusters. Understand the importance of topological spread in tolerating single fault domain failures and ensuring graceful handling of common zonal issues.