readme.md

lab-02 - build and deploy test application

Estimated completion time - 10 min

There is a simple C# rest API dotnet app, called GuineaPig that performs some CPU intensive computations, in order to simulate load in your cluster. The source code is located under src folder. If you use Visual Studio open app.sln file.

Goals

• build and push application image to Azure Container Registry
• deploy GuineaPig application into the cluster

Task #1 - build and push GuineaPig application

The simplest way to build and push image is to use az acr build command. Let's publish GuineaPig image to the Azure Container Registry.

Note! Don't forget to replace iacws5<YOU-UNIQUE-ID>acr with your ACR instance name.

# Go to aks-workshops\05-scaling-options-in-aks\src\GuineaPig folder
cd src\GuineaPig

# Build and publish guinea-pig:v1 image into your ACR
az acr build --registry iacws5<YOU-UNIQUE-ID>acr --image guinea-pig:v1 --file Dockerfile ..

Task #2 - deploy application to cluster

Create new deployment.yaml file with the following content. Replace image with your ACR instance name.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: guinea-pig
  labels:
    app: guinea-pig
spec:
  replicas: 1
  selector:
    matchLabels:
      app: guinea-pig
  template:
    metadata:
      labels:
        app: guinea-pig
    spec:
      containers:
      - name: api
        image: iacws5<YOU-UNIQUE-ID>acr.azurecr.io/guinea-pig:v1
        imagePullPolicy: IfNotPresent
        resources: 
          requests:
            cpu: 200m
          limits:
            cpu: 400m
        livenessProbe:
          httpGet:
            path: /health
            port: 80
          initialDelaySeconds: 3
          periodSeconds: 3    
---
apiVersion: v1
kind: Service
metadata:
  name: guinea-pig-service
  labels:
    app: guinea-pig
spec:
  ports:
  - port: 80
    protocol: TCP
    targetPort: 80
  selector:
    app: guinea-pig
  type: ClusterIP

Deploy application into default namespace.

# Deploy application into default namespace
kubectl apply -f deployment.yaml
deployment.apps/guinea-pig created
service/guinea-pig-service created

# Check that pod is up and running
kubectl get po
NAME                          READY   STATUS    RESTARTS   AGE
guinea-pig-6c994669b7-c8rtz   1/1     Running   0          32s

# Check that service was created
kubectl get svc
NAME                 TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)   AGE
guinea-pig-service   ClusterIP   10.0.49.164   <none>        80/TCP    47s

Task #3 - test application

Now let's test if application is actually up and running. We will use our old freind - busyboxplus:curl image, mainly because it contains curl command that we need for our testing.

# Run pod as interactive shell
kubectl run curl -i --tty --rm --restart=Never --image=radial/busyboxplus:curl -- sh

# Here is prompt from withing the pod
[ root@curl:/ ]$ 

# Call api endpoint. It should response with "[guinea-pig] - OK."
[ root@curl:/ ]$ curl http://guinea-pig-service/api
[guinea-pig] - OK.

In the separate terminal window, watch application logs.

# Get pod name
kubectl get po
NAME                          READY   STATUS    RESTARTS   AGE
guinea-pig-75f86bcf55-bzb5g   1/1     Running   0          2m18s

# Get guinea-pig application logs
kubectl logs guinea-pig-75f86bcf55-bzb5g -f
[21:26:36 WRN] Failed to determine the https port for redirect.
[21:26:56 INF] [guinea-pig] - OK.

Now let's test another application endpoint from the busybox shell.

[ root@curl:/ ]$ curl http://guinea-pig-service/api/highcpu
[ root@curl:/ ]$ curl http://guinea-pig-service/api/highcpu

You should see new logs appear from kubectl logs window

kubectl logs guinea-pig-75f86bcf55-bzb5g -f
[21:26:36 WRN] Failed to determine the https port for redirect.
[21:26:56 INF] [guinea-pig] - OK.
[21:26:58 INF] [guinea-pig] - OK.
[21:32:05 INF] [guinea-pig.highcpu] - execution took 18 ms.
[21:32:05 INF] [guinea-pig.highcpu] - execution took 18 ms.

Useful links

• Visual Studio 2019 Community Edition
• Download .NET 5.0
• Create your first Docker container with an ASP.NET web app
• Visual Studio Container Tools with ASP.NET Core
• Container Tools in Visual Studio
• How to configure Visual Studio Container Tools
• az acr build command
• Push your first image to a private Docker container registry using the Docker CLI

Next: create monitoring dashboard for test application

Go to lab-03