Kubernetes promises to run containerised workloads efficiently and at scale. It also provides auto-scaling features where the workloads can scale to meet the demands of the traffic. None of this is available by default and work needs to be done to reach the Kubernetes promised land. This talk will focus on how to architect Kubernetes clusters for high traffic websites or workloads. Attendees will learn how to configure ingress controller to securely expose workloads to external traffic. A demo will be presented which will show how Prometheus can be used to collect metrics from workloads. These metrics will then be fed to Kubernetes Pod Autoscaler using KEDA (Kubernetes Event-driven Autoscaling) to automatically scale the pods to meet the demand.