Container Orchestration

Container Orchestration

Container Orchestration

Containers solved one problem. Applications now ran consistently everywhere. But success created another challenge. A modern application doesn't run one container. It may run hundreds. Or even thousands. Some containers fail. Some need updates. Some need more CPU. Some need to communicate with each other. Managing them manually quickly became impossible. The engineering problem became clear. How do we deploy, scale, monitor and recover thousands of containers automatically? The engineering concept that solved this problem is Container Orchestration. An orchestrator continuously watches every container. If one fails, it starts another. If demand increases, it creates more containers. If demand falls, it removes the extras. It also decides where each container should run and keeps the entire application healthy without human intervention. Amazon Web Services provides this through Elastic Kubernetes Service (EKS) and Elastic Container Service (ECS). Microsoft Azure provides Azure Kubernetes Service (AKS). Google Cloud provides Google Kubernetes Engine (GKE). Different names. One engineering concept. Applications could now run at massive scale while largely managing themselves. But another challenge soon appeared. As applications grew, they exposed dozens or even hundreds of APIs. How do we provide one secure entry point for all those APIs?