It started as a routine deployment — a backend service on our AKS cluster, just another namespace, just another Ingress. But one TLS secret later, our pods were stuck, our PodDisruptionBudget was broken, and production was at risk.
In this talk, I’ll share a real-world debugging journey where a small change in a Kubernetes Ingress triggered cascading effects across namespaces and violated a PodDisruptionBudget. I’ll walk through how we diagnosed the issue, cordoned and uncordoned nodes to regain control, and leveraged replication factors for recovery.
Attendees will walk away understanding how PodDisruptionBudgets really behave under pressure, how node cordoning helps restore balance, and why even minor configuration changes can have major operational impacts.