Skip to content

Latest commit

 

History

History
73 lines (50 loc) · 4.01 KB

readme.md

File metadata and controls

73 lines (50 loc) · 4.01 KB
Raw

lab-07 - use cluster autoscaler to automatically scale an AKS cluster to meet application demands

To respond to changing pod demands, Kubernetes has a cluster autoscaler, that adjusts the number of nodes based on the requested compute resources in the node pool. By default, the cluster autoscaler checks the Metrics API server every 10 seconds for any required changes in node count. If the cluster autoscale determines that a change is required, the number of nodes in your AKS cluster is increased or decreased accordingly.

Cluster autoscaler is typically used alongside the horizontal pod autoscaler (HPA). When combined, the horizontal pod autoscaler increases or decreases the number of pods based on application demand, and the cluster autoscaler adjusts the number of nodes as needed to run those additional pods accordingly.

Task #1 - enable the cluster autoscaler

Since we already have a provision cluster with autoscaler disabled, we need to enable autoscaler first. Use the az aks update command with --enable-cluster-autoscaler parameter, and specify a node --min-count and --max-count.

# enable the cluster autoscaler
az aks update --resource-group iac-ws5-rg --name iac-ws5-aks --enable-cluster-autoscaler --min-count 1 --max-count 3

Task #2 - test autoscaler

Before we start, if your load-generator command is still running, stop it. If you can't find command, but your load-generator pod is still running, kill it.

# check if load-generator pos is up and running
kubectl get po
NAME                          READY   STATUS    RESTARTS   AGE
load-generator                1/1     Running   0          14m

# Kill load-generator pod
kubectl delete po load-generator
pod "load-generator" deleted

Next, let's delete HPA, because we want manually scale up number of replicas.

# delete HPA
kubectl delete hpa guinea-pig

If you use Windows Terminal I recommend you the following setup:

  • Left window for executing commands
  • Top right window to watch status changes with the pods - kubectl get po -w
  • Bottom right window to watch status changes with nodes - kubectl get nodes -w

pic

# Scale guinea-pig application up to 20 replicas
kubectl scale deployment/guinea-pig --replicas=20

You will see that Kubernetes will start scaling guinea-pg application up, will try to create 20 replicas, but you will notice that a lot of pods will be in Pending state. This is because there is not enough recurses available to schedule pods.

In about 1-2 mins, you should see two more nodes created. They will first have NotReady state and then eventually switch into Ready state.

We can use the technique to scale the number of nodes down.

# Scale guinea-pig application down to two replicas
kubectl scale deployment/guinea-pig --replicas=2

Observe your monitoring windows. First, you will see that Kubernetes will terminate all but two guinea-pg pods. Then it will take some time (sometimes more than 5 min)before AKS will start draining nodes. You will notice it when nodes status will change to NotReady. After that nodes will be decommissioned.

Useful links

Next: deploy KEDA

Go to lab-08