Coffee Application

Creating Deployments and Services via YAML configuration files

$ kubectl apply -f coffee-app.yaml
$ kubectl apply -f order-api.yaml

Access Kubernetes Dashboard

$ kubectl proxy &
// port defaults to 8001
// add -p [port-number] to specify which port to run the proxy
// e.g. kubectl proxy -p 8080 &

Access the Kubernetes Dashboard on http://localhost:8001/ui

Exposing and Accessing Application

The application is exposed through the service configuration in the yaml files.
There are 2 ways to access the application.

• Access via NodePort You will need your worker IP address and NodePort of the service

  Worker IP address:

  $ bx cs workers [cluster-name]
  OK
  ID                                                 Public IP      Private IP       Machine Type   State    Status   Version   
  kube-dal10-crbbdb1ff6a36846e9b2dfb522a07005af-w1   169.60.123.123   10.177.184.196   b1c.16x64      normal   Ready    1.7.4_1502*   
  

  NodePort:

  $ kubectl get svc
  NAME         CLUSTER-IP       EXTERNAL-IP     PORT(S)          AGE
  coffee-app   172.21.108.170   169.46.321.321   3000:30080/TCP   1h
  order-api    172.21.57.151    169.46.322.322   3001:30081/TCP   39m
  

  You will access the sample app through http://169.60.123.123:30080

• Access via LoadBalancer

  You can get the external IP through:

  $ kubectl get svc
  NAME         CLUSTER-IP       EXTERNAL-IP     PORT(S)          AGE
  coffee-app   172.21.108.170   169.46.321.321   3000:30080/TCP   1h
  order-api    172.21.57.151    169.46.322.322   3001:30081/TCP   39m
  

  You will access the sample app through http://169.60.321.321:3000

Scaling the Deployment

Scaling the coffee-app frontend

$ kubectl scale deploy coffee-app --replicas=2
kubectl get deployment "coffee-app" scaled

Scaling the order-api backend

$ kubectl scale deploy order-api --replicas=2
kubectl get deployment "order-api" scaled

You can scale up or down by specifying the number of replicas.

Horizontal Pod Autoscaler

$ kubectl autoscale deployment coffee-app --cpu-percent=20 --min=1 --max=10
$ kubectl autoscale deployment order-api --cpu-percent=20 --min=1 --max=10

// Send load to the backend
//Ctrl + C to cancel
while true; do curl -X POST -H 'Content-Type: application/json' -d '{ "mocha": "1", "americano": "2", "espresso": "3", "latte": "4", "machhiato": "5" }' http://169.46.322.322:3001/process_order; done

// Send load to the frontend
//Ctrl + C to cancel
while true; do curl -s http://169.46.74.114:3000 > /dev/null; done

// Send load to frontend then ordering
//Ctrl + C to cancel
while true; do curl -X POST -H 'Content-Type: application/json' -d '{ "mocha": "1", "americano": "2", "espresso": "3", "latte": "4", "machhiato": "5" }' http://169.46.74.116:3000/order; done

Debugging the Application

To view the logs in a pod

$ kubectl logs [pod-name]
or if you have multiple containers in a pod
$ kubectl logs [pod-name] -c [container-name]

To execute a command in a pod

$ kubectl exec [options] [pod-name]
e.g. kubectl exec -ti coffee-app-1522168388-6v3nj /bin/sh
i option allows stdin to the container and t option makes it a tty
This will execute /bin/sh in the container

Deletion of resources

$ kubectl delete -f coffee-app.yaml
$ kubectl delete -f order-api.yaml
$ kubectl delete hpa coffee-app order-api