Skip to main content
We’re almost at the end of this module — great work making it this far. As you start using containers on Azure, a common question arises: is AKS the only option, and is it the right platform for every workload? The short answer: it depends on how much control, flexibility, and operational responsibility your applications require. Below is a concise guide to the main Azure container hosting options, when to choose each, and the trade-offs to consider.

​
When you need full control: AKS, Kubernetes on VMs, Azure Red Hat OpenShift

  • AKS, self-managed Kubernetes on VMs, and Azure Red Hat OpenShift provide a full Kubernetes/OpenShift cluster experience.
  • These options expose cluster APIs so you can install operators and CRDs, run custom controllers, customize networking (custom CNI), and tune ingress, storage, and scheduling.
  • Choose these when you require:
    • Fine-grained orchestration control
    • Custom networking or ingress setups
    • Operator-managed software stacks or custom controllers
    • Advanced multi-tenant or hybrid scenarios
AKS supports both Linux and Windows containers via node pools. Windows containers are supported but have platform-specific constraints and fewer features compared to Linux node pools — factor this into planning and architecture decisions.
Self-managed Kubernetes and managed OpenShift (ARO) typically increase operational overhead and may incur additional licensing or management costs. Choose them only when you need the extra flexibility they provide.

​
Developer-focused managed platforms

  • Azure Container Apps
    • Built on Kubernetes but provides a higher-level, developer-centric experience.
    • The platform manages the underlying cluster; developers work with services, revisions, and DAPR-style service invocation instead of cluster APIs.
    • Built-in autoscaling (KEDA), traffic splitting, and observability make it excellent for microservices where you want rapid delivery with minimal ops.
  • Azure Container Instances (ACI)
    • Runs containers as single, isolated instances with strong container-level isolation.
    • No VM or cluster management — ideal for short-lived jobs, burst capacity, or isolated test environments.
    • ACI can be integrated as virtual nodes into AKS for burst scenarios.
  • Azure App Service
    • A PaaS for web apps and APIs with platform-managed runtimes or custom container support.
    • Two models: the “code” model (use platform runtime images) and the “container” model (Web App for Containers).
    • Good when you want a platform-managed web host with integrated deployment, authentication, and scaling.
  • Azure Functions
    • Event-driven, serverless compute for short-lived or event-triggered code.
    • Supports the Functions programming model, and can be deployed in containers when needed.
    • Use for lightweight, highly event-driven workloads with automatic scale-to-zero and bindings.

​
Quick comparison table

Resource TypeBest forAccess & controlScaling behaviorTypical use cases
AKS / K8s on VMs / AROFull orchestration controlFull cluster API, CRDs, operatorsUser-defined (cluster autoscaler, HPA, custom)Complex microservices, operator-driven apps, custom networking
Azure Container AppsManaged microservicesNo direct K8s API; platform-managedBuilt-in KEDA autoscalingEvent-driven microservices, APIs, service-to-service apps
Azure Container Instances (ACI)Short-lived or isolated containersContainer-level only; no cluster controlPer-container allocation (fast start)Jobs, CI/CD tasks, ephemeral testing, burst capacity
Azure App ServiceManaged web apps/APIsPlatform runtimes or custom containersPlatform-managed scalingWeb apps, APIs, straightforward containerized apps
Azure FunctionsEvent-driven, serverlessFunction model, can run in containersAutomatic, scale-to-zeroEvent handlers, cron jobs, serverless APIs

​
Decision guidance — pick by workload needs

  • If you need maximum control, advanced networking, or operators: choose AKS, ARO, or self-managed Kubernetes.
  • If you want Kubernetes benefits without operating a cluster: choose Azure Container Apps.
  • For single-container, short-lived tasks or fast-start isolated runs: choose ACI.
  • For traditional web apps with minimal infra management: choose App Service (code model or container).
  • For event-driven, short-duration logic with minimal ops: choose Azure Functions.
An infographic that maps Azure container hosting options on two axes — from infrastructure management to serverless and from unopinionated to opinionated containers. It shows examples like Azure Kubernetes Service, Kubernetes on VMs, Azure Red Hat OpenShift, Azure Container Instances and Container Apps, plus Azure App Service and Azure Functions.
Consider the trade-offs between control and operational overhead, and choose the service that best matches your application’s needs for customization, isolation, scaling, and developer experience. Review the links above and test a small proof-of-concept in the chosen platform before committing to it in production.

Watch Video