The second approach for running cloud Kubernetes on Azure is to use Azure Kubernetes Service, or AKS, Azure’s managed Kubernetes service. AKS is similar to EKS in that it automates most of the work required to manage the underlying infrastructure, as well as the Kubernetes control plane.
To create a Kubernetes cluster in AKS, run a command like the following using az, the CLI tool for Azure:
az aks create \ --resource-group myResourceGroup \ --name myAKSCluster \ --node-count 2 \ --generate-ssh-keys \ --attach-acr <acrName>As with EKS, the cluster creation process can take several minutes. Once your cluster is up and running, you can begin interacting with it using kubectl.
Overall, AKS is simpler than EKS, partly because it doesn’t have as many optional features. On AKS, there are no nuances equivalent to nodegroups or Fargate mode that you have to contend with.
On the other hand, AKS is less scalable than EKS. The maximum number of nodes you can run in AKS is 1,000, compared to 3,000 for EKS. That limit probably won’t matter for most use cases, but it is important to consider if you plan a very large deployment.
EKS and AKS pricing is in general similar; however, AKS can be less expensive because you don’t necessarily have to pay a per-hour fee for your clusters. Such a fee exists in AKS (and it happens currently to be the same price - $0.10 per cluster per hour – as EKS’s fee), but you can opt out of this fee in ASK by choosing to forgo the uptime guarantees that are built into AKS. So, if you want to reduce your AKS costs, you can agree to a lower level of reliability. EKS has no such option.