Waseem Shaikh

Kubedoom DevOps Project

Kube DOOM Project for Kubernetes learning purposes


Kill Kubernetes pods using Id's Doom!


The next level of chaos engineering is here! Kill pods inside your Kubernetes cluster by shooting them in Doom!

Youtube Video: https://youtu.be/pHLb3GMyNhI

GitHub Repo : https://github.com/waseemuddin/kubedoom-project-devops

This repo contains about the Kube Doom project for learning and implementing the Kubernetes and killing the pods inside your kubernetes cluster.

The credit goes to iiDKx & storax/kubedoom.

The following steps are involved to implement the KubeDoom Project



This is project can be implemented on your local machine (VM), AWS Cloud or any other cloud provider which suites you.

kubedoom Diagram



Step 1: Update VM/Ubuntu


After installation of VM (ubuntu), please make sure your sytstem is updated

#Basic command to make sure all are update
sudo apt update
sudo apt-get update
#Check IP
ip --brief addr show
#Check firewall
sudo ufw status
sudo ufw allow 22/tcp
sudo -i
swapoff - a

Step 2: Steup Kubernetes Cluster (Kind)


Running Kubedoom inside Kubernetes

Install kind master and worker node

See the example in the /manifest directory. You can quickly test it using kind. Create a cluster with the example config from this repository:

$ kind create cluster --config kind-config.yaml
Creating cluster "kind" ...
✓ Ensuring node image (kindest/node:v1.25.3) 🖼
✓ Preparing nodes 📦 📦
✓ Writing configuration 📜
✓ Starting control-plane 🕹️
✓ Installing CNI 🔌
✓ Installing StorageClass 💾
✓ Joining worker nodes 🚜
Set kubectl context to "kind-kind"
You can now use your cluster with:

kubectl cluster-info --context kind-kind
Not sure what to do next? 😅 Check out https://kind.sigs.k8s.io/docs/user/quick-start/

Kind Master and worker node


Set your Kube context with: kubectl cluster-info --context kind-kind

This will spin up a 2 node cluster inside docker, with port 5900 exposed from the worker node. Then run kubedoom inside the cluster by applying the manifest provided in this repository. The Kubernetes-Manifests-Files directory holds Kubernetes manifests for deploying your application on cluster.

$ kubectl apply -k manifest/
namespace/kubedoom created
deployment.apps/kubedoom created
serviceaccount/kubedoom created
clusterrolebinding.rbac.authorization.k8s.io/kubedoom created

kubedoom namespace


Step 3: Docker Build


Build the image with docker build --build-arg=TARGETARCH=amd64 . -t kubedoom while in this directory. Then run:

kubedoom namespace


#docker run command and and attached pods
$ docker run -p5801:5800 \
-e NAMESPACE=default \
--net=host \
-v ~/.kube:/root/.kube \
--rm -it --name kubedoom \

Optionally, if you set -e NAMESPACE={your namespace} you can limit Kubedoom to deleting pods in a single namespace

kubedoom namespace


Step 4: Docker Build


Deploy some Nginx pods to your cluster:

kubectl apply -f k8s/nginx-deployment.yaml

You can refresh the dashboard as you kill the pods from Kubedoom and expand upon this to track other metrics and applications in your cluster.

kubedoom namespace


Step 5: Attaching a VNC Client


Now start a VNC viewer and connect to localhost:5900. The password is idbehold:

$ vncviewer viewer localhost:5901

You should now see DOOM! Now if you want to get the job done quickly enter the cheat idspispopd and walk through the wall on your right. You should be greeted by your pods as little pink monsters. Press CTRL to fire. If the pistol is not your thing, cheat with idkfa and press 5 for a nice surprise. Pause the game with ESCiddqd for god mode.

Cheat codes found here: https://doom.fandom.com/wiki/Doom_Cheat_Codes

kubedoom namespace


kubedoom namespace


To connect run:
$ vncviewer viewer localhost:5900

Step 6: Kubedoom demo




kubedoom namespace


Create a dashboard in Grafana to monitor the Nginx containers. To do this open Grafana at http://localhost:3000 and login. Make sure you have your data source set to your Prometheus pod from the previous step. From the left hand menu, create a new dashboard and add a panel. Select the panels dropdown menu, select Inspect and then select Panel JSON. Here you will be able to delete the current JSON and replace it with JSON from the grafana folder ./grafana/nginx-panel.json. Save and apply this and you should be able to see the CPU usage of the current deployed pods.

Step 7: Deploying Prometheus


Create a monitoring namespace to keep things tidy.

kubectl create namespace monitoring

Deploy Prometheus to scrape and store metrics for your cluster with:

kubectl apply -f k8s/prometheus.yaml -n monitoring

List the Prometheus pod name and IP address.

$ kubectl get pods -o wide -n monitoring
prometheus-deployment-75cff7d89f-w422q 1/1 Running 1 (15m ago) 25m kind-worker <none> <none>

In a separate terminal run the below command to port-forward and you'll be able to access Prometheus on http://localhost:8080:

kubectl port-forward -n monitoring prometheus-deployment-75cff7d89f-w422q 8080:9090

To run in background:

kubectl port-forward -n monitoring prometheus-deployment-75cff7d89f-w422q 8080:9090 &

Prometheus graph display command.

rate(container_cpu_usage_seconds_total{namespace="default", container="nginx"}[30s]) * 100

Step 8: Deploying Grafana


Deploy Grafana to graph our metrics from Prometheus with:

kubectl apply -f k8s/grafana.yaml -n monitoring

In another terminal, list the pod names and port-forward from one of the pods:

$ kubectl get pods -n monitoring
grafana-5469c64c7d-ddz4r 1/1 Running 1 (20m ago) 30m
grafana-5469c64c7d-xdlmw 1/1 Running 1 (20m ago) 30m
prometheus-deployment-75cff7d89f-w422q 1/1 Running 1 (20m ago) 30m

$ kubectl port-forward -n monitoring grafana-5469c64c7d-ddz4r 3000
Forwarding from -> 3000
Forwarding from [::1]:3000 -> 3000

Grafana should now be reachable at http://localhost:3000.

Log in with username admin and password admin.

Once you are logged in, you will need to go to the Settings (gear icon bottom left) and edit the Data sources.

Change the URL in the settings from http://prometheus-service.monitoring.svc:8080 to http://<Prometheus Pod IP>:9090. Use the below command to get the Prometheus pod IP.

$ kubectl get pods -o wide -n monitoring
prometheus-deployment-75cff7d89f-w422q 1/1 Running 1 (15m ago) 25m kind-worker <none> <none>

So I would enter as my data source. Save and test this.

Kubedoom requires a service account with permissions to list all pods and delete them and uses kubectl 1.25.3.

Running Locally


In order to run locally you will need to

  1. Run the kubedoom container
  2. Attach a VNC client to the appropriate port (5901)

With Podman


Run kubedoom:latest with podman locally:

$ podman run -it -p5901:5900/tcp \
-v ~/.kube:/tmp/.kube --security-opt label=disable \
--env "KUBECONFIG=/tmp/.kube/config" --name kubedoom

Killing namespaces


Kubedoom now also supports killing namespaces in case you have too many of them. Simply set the -mode flag to namespaces:

$ docker run -p5901:5900 \
--net=host \
-v ~/.kube:/root/.kube \
--rm -it --name kubedoom \
kubedoom:latest \
-mode namespaces

Building Kubedoom


The repository contains a Dockerfile to build the kubedoom image. You have to specify your systems architecture as the TARGETARCH build argument. For example amd64 or arm64.

$ docker build --build-arg=TARGETARCH=amd64 -t kubedoom .

To change the default VNC password, use --build-arg=VNCPASSWORD=differentpw.

Leave a Comment