Azure Kubernetes Service (AKS)

Azure Kubernetes Service (AKS) is Microsoft's managed Kubernetes service for teams on Azure, run in Automatic or Standard mode and priced per cluster.

AKS takes the job of building and patching a Kubernetes control plane away from the team that ships the containers. Azure operates the kube-apiserver, etcd and kube-scheduler, and the team deploys its applications onto node pools of Azure virtual machines. The service is part of Microsoft's Azure cloud, with one portal and one documentation set for every cluster. It is CNCF-certified, so it serves the upstream Kubernetes API that standard tooling expects. AKS Automatic preconfigures node provisioning, scaling, upgrades and network settings, and manages the nodes itself. AKS Standard keeps the managed control plane but leaves node pools and their settings to the operator. Node auto-repair reimages or redeploys an unhealthy node, and the cluster autoscaler adds or removes nodes as demand changes. Pods scale on CPU, memory, application metrics or queue backlogs through horizontal and vertical pod autoscaling and the KEDA add-on. The application routing add-on supplies managed NGINX ingress, and the Secrets Store CSI driver mounts Key Vault secrets into pods. Platform and DevOps teams that already run on Azure are the core users, above all those who want Kubernetes without running its control plane. AKS Automatic suits teams that want production-ready defaults and little node tuning. AKS Standard suits operators who need control over node pools and the virtual machine types behind them. Its SOC, ISO, PCI DSS and HIPAA compliance matters to organisations that handle payment or health data. Teams with devices in the field can take the same Kubernetes model to the edge with AKS Edge Essentials. A typical deployment runs microservices from images in Azure Container Registry, split across namespaces with their own resource quotas. Access goes through Microsoft Entra ID, which ties Kubernetes RBAC roles to the organisation's directory. Cilium, Calico or Azure Network Policy Manager limits which workloads can talk to each other on the cluster network. Azure Monitor, Azure Policy and Defender for Containers watch, govern and protect the same clusters. Teams practising GitOps add the Flux extension, which keeps each cluster in step with the manifests in a Git repository. Cluster auto-upgrade moves the control plane first and then each node pool to a new Kubernetes version, which keeps long-running production clusters current.