Horizontal scaling adjusts application capacity by adding or removing application instances as workload changes. Kubernetes replicates the application so developers do not manually install, configure, or run each copy. section 1.2.1 section 3.3.3
Kubernetes scales Pods, not individual containers. A Deployment maintains the desired number of Pods, while a Service provides one stable address and distributes requests across the available replicas. section 3.4 section 5.7
Scale a Deployment to three replicas
Run `kubectl scale deployment kiada --replicas=3`. Kubernetes creates the additional Pods; `kubectl get deploy` can then show `3/3` ready replicas. Requests sent to the Service reach different Pods in varying order, demonstrating internal load balancing.
Scaling remains the application's responsibility: Kubernetes makes replication easy but cannot make an application scalable by itself. Separating frontend and backend workloads into different Pods allows each part to scale independently. section 3.3.3 section 5.7
