This is the sample project to use the docker(swarm) with kubernetes. The python app has been built as a docker image and pushed to AWS ECR repository. Using the docker stack we are deploying it in kubernetes as service.

1. Tools
   1. docker desktop installed
   2. Kubernetes enabled in docker desktop
   3. AWS ECR as docker registry
2. Steps
   1. Build/Tag/Push docker image
   2. Create docker compose file
   3. Create deployment
   4. Scale deployment
   5. Delete deployment

Build/ Tag/ Push docker image

• To build
  • docker build -t apiv1.2 .
• to Tag to ECR
  • docker tag apiv1.2 911456480347.dkr.ecr.ap-south-1.amazonaws.com/udacity:1.2
• To push to ECR
  • docker push 911456480347.dkr.ecr.ap-south-1.amazonaws.com/udacity:1.2

Creating docker compose file

We created a simple docker file which has a service and its ports. We can add more attributes like replica sets /volumes/ networks to this service resource in the docker file.

Creating the kubernetes deployment

Create a docker stack using the docker compose file. Please do not forget to mention the orchestrator name as kubernetes. The default is swarm orchestrator.

```docker stack deploy --orchestrator kubernetes --compose-file .\docker-compose.yml hello-py-api```

Removing the kubernetes service deployment

`docker stack rm --orchestrator kubernetes hello-py-api`

Scaling pods

To scale up we can update the deployment using docker command or use kubectl to update the replicas .

• get the deployment names

  kubectl get deployments

• scale up

  kubectl scale deployments/api --replicas=5

• Check pods status

  kubectl get pods -o wide

• Scaling down

  kubectl scale deployments/api --replicas=2

• Check pods status

  kubectl get pods -o wide

Delete the deployment

• To list all the deployments in the stack docker stack --orchestrator kubernetes ls
• To delete an deployment docker stack --orchestrator kubernetes rm hello-py-api
• get the deployment names kubectl get deployments