AKS Overview
What is AKS?
Azure Kubernetes Service (AKS) is Microsoft’s managed Kubernetes offering. It abstracts away much of the complexity of cluster setup and maintenance, allowing teams to focus on deploying and managing containerized applications. AKS handles tasks like provisioning, scaling, upgrading, and integrating with Azure services, making it a powerful platform for running production workloads in the cloud.
Why Use AKS?
Building Kubernetes clusters from scratch is hard. Managing and operating them into production is even harder. AKS simplifies Kubernetes operations by:
- Automating cluster provisioning and upgrades
- Providing a fully managed control plane
- Enabling high availability through availability zones and multi-region strategies
- Being backed by Microsoft support and SLAs
- Offering built-in monitoring, diagnostics and self healing
- Supporting autoscaling and multiple node pools
- Integrating deeply with Azure networking, storage, and security services
- Supporting secure access via Entra ID
- Seamlessly integrating with Azure Container Registry
AKS Tiers
AKS is available in three tiers. All three tiers offer the same core features, but have differing supported cluster sizes and SLAs. For production workloads, a minimum of Standard tier is strongly recommended. The Free tier is suitable for development and testing purposes.
Tier | Description | Good For | Supported Cluster Types | Pricing | Key Features |
---|---|---|---|---|---|
Free | Entry-level tier with no SLA, ideal for learning and testing | Learning, experimentation, dev/test | Up to 1,000 nodes | Free cluster management; Pay-as-you-go for compute | All AKS features; No SLA; Basic support only |
Standard | Production-ready tier with SLA and full feature support | Production workloads, general use | Up to 5,000 nodes | Cluster management billed per hour; Pay-as-you-go for compute | 99.9–99.95% SLA; Uptime SLA enabled; Full support |
Premium | Enhanced support and reliability for mission-critical apps | Regulated industries, mission-critical apps | Up to 5,000 nodes | Higher cluster management cost; Pay-as-you-go for compute | Extended support lifecycle; All Standard features; Advanced reliability |