TheCloudBootcamp Day 2

Day 2 of the challenge started off with a bang. Working with source code for the fictious comapny “CloudMart”, the goal for the day was to build a containers for the front and backend of the app and deploy on AWS EKS. Kubernetes is so much fun, thrilled to see the scenario use it.

Since this was the first pass of getting everything working we will be staring out with manual builds to make sure everything works and we can mature the build process down the line.

Now that we had the source code I set out to build the container images and created Dockerfiles for each. The images were built and stored in AWS Elastic Container Repository (ECR). Now that we have the images we could setup a cluster in AWS Elsatic Kubernetes Service (EKS). Storing the images in ECR made it simple to launch on EKS.

I used the following Kubernetes deployment file, created in YAML, for the backend:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: cloudmart-backend-app
spec:
  replicas: 1
  selector:
    matchLabels:
      app: cloudmart-backend-app
  template:
    metadata:
      labels:
        app: cloudmart-backend-app
    spec:
      serviceAccountName: cloudmart-pod-execution-role
      containers:
      - name: cloudmart-backend-app
        image: public.ecr.aws/<insert repository id here>/cloudmart-backend:latest
        env:
        - name: PORT
          value: "5000"
        - name: AWS_REGION
          value: "us-east-1"
        - name: BEDROCK_AGENT_ID
          value: "xxxxxx"
        - name: BEDROCK_AGENT_ALIAS_ID
          value: "xxxx"
        - name: OPENAI_API_KEY
          value: "xxxxxx"
        - name: OPENAI_ASSISTANT_ID
          value: "xxxx"
---

apiVersion: v1
kind: Service
metadata:
  name: cloudmart-backend-app-service
spec:
  type: LoadBalancer
  selector:
    app: cloudmart-backend-app
  ports:
    - protocol: TCP
      port: 5000
      targetPort: 5000

We added a few lines for Bedrock and OpenAI for later in the challenge to reduce the changes later on in the challenge.

With the EKS cluster and the YAML deplyemnet file it was easy to deploy the backend with the a simple command:

kubectl apply -f cloudmart-backend.yaml

To check the deployment I used the following commands:

kubectl get pods
kubectl get deployment
kubectl get service

It was a rinse and repeat step for the front end except that the font end needed a quick .env file to point the URL of the backend service so the app can pull data. This URL needed is the output of the kubectl get service.

And with the a quick deployment command:

kubectl apply -f cloudmart-frontend.yaml

We had a fully deployed application runnning on AWS EKS.

Let’s see what Day 3 holds in store!!!!