Resource Limits

In this training, you will learn about resource limits.

The application implements a CPU and a Memory leak.

Change into the lab directory:

cd /workspaces/kubernetes-fundamentals-for-devs/07_resources

Create the Pod

Inspect pod.yaml and service.yaml definition file and create the pod and the service

cat k8s/pod.yaml
kubectl create -f k8s/pod.yaml

Leaking CPU

Learn what happens when your application reaches its CPU limits.

# [TERMINAL-2] Attach to the application
kubectl attach -it my-app

# [TERMINAL-2] Engage the CPU leak
leak cpu

# Check what is happening with the Pod.
watch -n 1 kubectl top pods

Important

Note that the Container does not get restarted. The amount of CPU is limited to 30 MilliCores.

Leaking Memory

Learn what happens when your application reaches its Memory limits.

# Restart the Pod
kubectl delete pod my-app --force --grace-period=0
kubectl apply -f k8s/pod.yaml

# [TERMINAL-2] Attach to the application
kubectl attach -it my-app

# [TERMINAL-2] Engage the Memory Leak
leak mem

# After the Container gets restarted (~ 10 seconds) you see this in the RESTARTS column of
kubectl get pods

# The reason for the last restart (=OOMKilled) you can find out via the following command
kubectl get pod my-app -o jsonpath='{.status.containerStatuses[0].lastState}' | jq

Cleanup

Delete the created resources.

kubectl delete -f k8s/ --force --grace-period=0