I am always on the lookout for cool and interesting demos to deploy, especially with some of the work I have been doing lately with vSphere with Kubernetes (K8s) and Tanzu Kubernetes Grid (TKG). I am sure many of you have probably seen the basic wordpress demos which seems to be the typical "Hello World" app for K8s and having something more compelling not only makes the demo more interesting but it can also help folks better understand how a modern applications can be built, deployed and run.
Below is a list of of the K8s demo applications that I have come across as part of my exploration and by no means is this an exhaustive list. I have been able to successfully deploy these applications running on the latest version of K8s (1.17 and 1.18) as I did come across other demos which did not work or I had issues setting up. If there are other K8s demos that folks have used, feel free to leave a comment and I will update the blog post after doing some basic testing.
For those of you who may not have a K8s environment and is running either vSphere 6.7 Update 3 or have access to a VMware Cloud on AWS SDDC, you can easily setup a TKG Cluster in under 30 minutes leveraging my TKG Demo Appliance Fling.
Yelb
Yelb has been one of my go to demos due to its simplicity. For those not familiar, Yelb is a famous VMware demo application that was built by my good friend Massimo Re Ferre' and Andrea Siviero. This web application contains the following services: UI Frontend, Application Server, Database Server and Caching Service using Redis.
Special Requirements:
- None
NodePort Deployment:
Step 1 - Deploy the application
kubectl create ns yelb
kubectl apply -f https://raw.githubusercontent.com/lamw/vmware-k8s-app-demo/master/yelb.yaml
Step 2 - Verify all pods are ready
kubectl -n yelb get pods
Step 3 - To access the application, open web browser to http://<ip>:30001
kubectl -n yelb describe pod $(kubectl -n yelb get pods | grep yelb-ui | awk '{print $1}') | grep "Node:"
LoadBalancer Service Deployment:
Step 1 - Deploy the application
kubectl create ns yelb
kubectl apply -f https://raw.githubusercontent.com/lamw/vmware-k8s-app-demo/master/yelb-lb.yaml
Step 2 - Verify all pods are ready
kubectl -n yelb get pods
Step 3 - To access the application, open web browser to http://<external-ip>
kubectl -n yelb get svc/yelb-ui
Note: It is expected that your K8s cluster can support a LoadBalancer Service
VMware Event Broker Application
I know many of you are already familiar with the VMware Event Broker Appliance (VEBA) Fling which enables customers to easily build event-driven automation based on vCenter Server Events. Since VEBA is built as a native K8s application, customers that have an existing K8s Cluster which includes vSphere with K8s and TKG can also get the benefits of VEBA!
Special Requirements:
- None
Deployment:
Detailed instructions on deploying and configuring VEBA with K8s can be found here.
Kubeapps
I recently learned about Kubeapps which provides a very slick web interface for deploying and managing K8s applications which also includes Helm charts. After deploying Kubeapps, you immediately have access to a catalog of several hundred ready to deploy applications. It includes basic apps like WordPress but it also contains some of the most popular and most up to date open source applications that are being deployed in production like Kafka, Jenkins and Elastic Search as an example. If you have your own application repository, you can add those to Kubeapps which will automatically give you UI interface to deploy your own custom K8s applications. Kubeapps is also a VMware Opensource project that is being led by the Bitnami team which was acquired recently by VMware. Definitely worth checking out!
Special Requirements:
- Helm is required for installation
- Storage Classes (for persistent apps)
Deployment:
Detailed instructions can be found here and to learn more about Kubeapps, Tiffany Jernigan, a Developer Advocate for VMware has fantastic getting started video which I also recommend checking out.
ACME Fitness
The ACME Fitness application is built by our VMware Cloud Advocacy team and it demonstrates an example e-commerce/online application which is made up of a number of different microservices. It also includes a load generator tool which is also looks pretty neat.
Special Requirements:
- Load Balancer required
Deployment:
Step 1 - Clone repo
git clone https://github.com/vmwarecloudadvocacy/acme_fitness_demo.git
cd acme_fitness_demo/kubernetes-manifests
Step 2 - Specify a secret name and namespace to apply to the deployment. If you do not have the permission to create a namespace, simply omit that from each of the commands.
ACME_SECRET=VMware1!
ACME_NAMESPACE=acme-fitness
Step 3 - Deploy application
kubectl create ns ${ACME_NAMESPACE}
kubectl -n ${ACME_NAMESPACE} create secret generic cart-redis-pass --from-literal=password=${ACME_SECRET}
kubectl -n ${ACME_NAMESPACE} apply -f cart-redis-total.yaml
kubectl -n ${ACME_NAMESPACE} apply -f cart-total.yaml
kubectl -n ${ACME_NAMESPACE} create secret generic catalog-mongo-pass --from-literal=password=${ACME_SECRET}
kubectl -n ${ACME_NAMESPACE} create -f catalog-db-initdb-configmap.yaml
kubectl -n ${ACME_NAMESPACE} apply -f catalog-db-total.yaml
kubectl -n ${ACME_NAMESPACE} apply -f catalog-total.yaml
kubectl -n ${ACME_NAMESPACE} apply -f payment-total.yaml
kubectl -n ${ACME_NAMESPACE} create secret generic order-postgres-pass --from-literal=password=${ACME_SECRET}
kubectl -n ${ACME_NAMESPACE} apply -f order-db-total.yaml
kubectl -n ${ACME_NAMESPACE} apply -f order-total.yaml
kubectl -n ${ACME_NAMESPACE} create secret generic users-mongo-pass --from-literal=password=${ACME_SECRET}
kubectl -n ${ACME_NAMESPACE} create secret generic users-redis-pass --from-literal=password=${ACME_SECRET}
kubectl -n ${ACME_NAMESPACE} create -f users-db-initdb-configmap.yaml
kubectl -n ${ACME_NAMESPACE} apply -f users-db-total.yaml
kubectl -n ${ACME_NAMESPACE} apply -f users-redis-total.yaml
kubectl -n ${ACME_NAMESPACE} apply -f users-total.yaml
kubectl -n ${ACME_NAMESPACE} apply -f frontend-total.yaml
kubectl -n ${ACME_NAMESPACE} apply -f point-of-sales-total.yaml
Step 4 - Verify all pods are in ready. I did notice that the point-of-sales (pos) pod may
kubectl -n ${ACME_NAMESPACE} get pods
Step 5 - To access the application, open web browser to http://<external-ip>
kubectl -n ${ACME_NAMESPACE} get svc/frontend
Online Boutique
Online Boutique is another e-commerce/online store demo that was built by the folks over at Google which demonstrates a 10-Tier microservice application.
Special Requirements:
- Load Balancer required
Step 1 - Clone repo
git clone https://github.com/GoogleCloudPlatform/microservices-demo.git
cd microservices-demo
Step 2 - Create namespace to apply to the deployment. If you do not have the permission to create a namespace, simply omit that from each of the commands.
kubectl create ns boutique
Step 3 - Deploy application
kubectl -n boutique apply -f ./release/kubernetes-manifests.yaml
Step 4 - Verify all pods are ready
kubectl -n boutique get pods
Step 5 - To access the application, open web browser to http://<external-ip>
kubectl -n boutique get svc/frontend-external
Robot Shop
Robot Shop is another e-commerce/online store application that was built by Instana using a number of different web technologies and microservices. This might be your cup of tea for those Sci-Fi fans 🙂
Special Requirements:
- Helm is required for installation
- Load Balancer required
Step 1 - Clone repo
git clone https://github.com/instana/robot-shop.git
cd robot-shop/K8s/helm
Step 2 - Create namespace to apply to the deployment. If you do not have the permission to create a namespace, simply omit that from each of the commands.
kubectl create ns robot-shop
Step 2 - Deploy application
kubectl create ns robot-shop
helm install robot-shop --namespace robot-shop .
Step 3 - Verify all pods are ready
kubectl -n robot-shop get pods
Step 4 - To access the application, open web browser to http://<external-ip>:8080
kubectl -n robot-shop get svc/web
Kubedoom
Kubedoom is a fun little application that I came across which allows you to have some fun killing demons, err I mean pods within your K8s Cluster to test the resiliency of an application. Its hard work but someone has to do it 😉
via GIPHY
Special Requirements:
- VNC Client
Step 1 - Clone repo
git clone https://github.com/storax/kubedoom.git
cd kubedoom
Step 2 - Deploy application
kubectl apply -f manifest/
Step 3 - Identify the IP Address of the K8s Node to connect
kubectl -n kubedoom describe pod | grep 'Node:'
Note: If you are using the TKG Demo Appliance, you can also setup SSH port forward to the K8s Node by using ssh [email protected][TKG-IP] -L 5900:[K8s-NODE-IP]:5900 and then connecting to localhost:5900 from your local desktop after successfully establishing SSH tunnel.
Step 4 - Connect to the Kubedoom container using any VNC Client on port 5900 and the password is 1234.
Kubevaders
I initially had Kubevaders on my list but was not able to get it to work. I finally got a chance to take another look and figured it out. If Doom is not your cup of tea, Kubevaders is another fun way of killing pods.
Special Requirements:
- Helm is required for installation
- Load Balancer required
Note: If you are using my TKG Demo Appliance, you can deploy Metallb for the Load Balancer by following the instructions here.
Step 1 - Clone repo
git clone https://github.com/lucky-sideburn/KubeInvaders.git
cd Kubevaders
Step 2 - Install an Ingress Controller. For simplicity purposes, we are using Nginx and will be installing via Helm:
helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx
helm install nginx-ingress ingress-nginx/ingress-nginx
Step 3 - Create namespace and deploy application using Helm. You will need to specify a list of K8s Namespaces where you want the Pods to be made available for your destruction. Here are am using the Acme Fitness application which has quite a few pods. You can specify more than one namespace, just make sure to escape the "comma" by doing something like "ns1\,ns2\,ns3"
kubectl create namespace kubeinvaders
helm install kubeinvaders --set-string target_namespace="acme-fitness" --namespace kubeinvaders ./helm-charts/kubeinvaders --set ingress.hostName=tkg-k8s-invaders.io
Step 4 - Retrieve the external IP Address that should have been allocated to your Nginx Ingress Controller
kubectl get svc/nginx-ingress-ingress-nginx-controller
Step 5 - As part of deploying the application, there was an ingress hostname property that was set which by default we used tkg-k8s-invaders.io which will be used to access the demo. You can either create a DNS entry or you can simply create a hosts entry that maps the hostname to the IP Address.
Step 6 - Open browser from a system that can access the external IP Address/Hostname and you should now be taken to Kubevaders application. You can switch between the different namespaces and start killing your pods!
Thanks for the comment!