Workshop
Requirements
In order to run the workshop you need to install the following tools:
- kubectl - https://kubernetes.io/docs/tasks/tools/
- kind - https://kind.sigs.k8s.io/docs/user/quick-start
- docker - https://docs.docker.com/engine/install
- make - package (sourceforge for Windows)
Those following packages are needed to spin the lab used during the workshop. We are reusing our developing environment.
For Mac user we have a oneliner to install everything (if you are using brew):
The last requirements is of course kubehound. You need to download the latest release from our repository:
wget https://github.com/DataDog/KubeHound/releases/latest/download/kubehound-$(uname -o | sed 's/GNU\///g')-$(uname -m) -O kubehound
chmod +x kubehound
or
Cheatsheet
Starting the lab
First you need to run spin our dev environement with a vulnerable cluster:
Initiating Kubehound
As the images used by KubeHound are quite heavy (due to Jupyter and Janusgraph), we want to make sure that we have them downloaded before starting the workshop. To do so, we can run the following command:
This will pull all the images that will be needed during the workshop.
Running the workshop
In order to use our vulnerable cluster, we need to use the kubeconfig
file generated when we created (with kind) our cluster. This variable needs to be exported in all the shell you will be using during the workshop.
export KUBECONFIG=./test/setup/.kube-config
# Checking the clustername
kubectl config current-context
# Checking the pods deployed
kubectl get pods
During the workshop we will be playing with Kubernetes resources. We advise you to install k9s which is a great tool made by the community - provides a terminal UI to interact with k8s cluster.
In order to test the attacks, we will assume breach of the containers. To execute a command you can jump into a container/pod using the following command:
Note
You can also use k9s (typing on s
key when highlighting a pod).