DNS is Simple. DNS is Hard (wespiser.com)

The article argues that DNS looks like simple domain-to-IP configuration, but at internet scale it operates like an unmanaged distributed system with recursive resolvers, caching, and weak visibility. It describes real outages—such as Dyn’s 2016 failure and an AWS/DynamoDB DNS-related incident—showing how inconsistent or stale DNS state can break dependent services for hours. It also recounts a provider migration where cached DNS state in Kubernetes prevented timely convergence, leading to manual restarts. Overall, it concludes that DNS changes require thinking in terms of distributed coordination and time-based convergence rather than instant cutovers.