You can manually scale nodes to test how your applications responds to a change in available resources and state. Manually scaling resources also lets you define a set amount of resources to use to maintain a fixed cost, such as the number of nodes. To manually scale, you define node count. The Kubernetes API then draining nodes based on node count.
When scaling down nodes, the Kubernetes API calls the relevant Azure Compute API tied to the compute type used by your cluster. For example, for clusters built on VM Scale Sets the logic for selecting which nodes to remove is determined by the VM Scale Sets API.
You can adjust the number of nodes manually if you plan more or fewer container workloads on your cluster.
The following command increases the number of nodes to two. The command takes a couple of minutes to complete.
# Check how many nodes are in the cluster now
kubectl get nodes
NAME STATUS ROLES AGE VERSION
aks-system-36621428-vmss000000 Ready agent 18m v1.21.7
# scale up cluster to two nodes
az aks scale --resource-group iac-ws5-rg --name iac-ws5-aks --node-count 2
# Get number of nodes
kubectl get nodes
NAME STATUS ROLES AGE VERSION
aks-system-36621428-vmss000000 Ready agent 19m v1.21.7
aks-system-36621428-vmss000001 Ready agent 43s v1.21.7
# scale cluster down to one node
az aks scale --resource-group iac-ws5-rg --name iac-ws5-aks --node-count 1
# Get number of nodes
kubectl get nodes
NAME STATUS ROLES AGE VERSION
aks-system-36621428-vmss000000 Ready agent 19m v1.21.7