CRUD application with FastAPI

Run in local

You can test the application in your local with docker.

docker-compose up -d --build

After containers successfully run, you can check an endpoint with curl command

curl -s localhost:8000/notes | jq -r
{
  "notes": [
    {
      "id": 1,
      "title": "Beyond the legacy code",
      "description": "Awesome book!"
    },
    {
      "id": 2,
      "title": "Effective Java",
      "description": "Difficult..."
    }
  ]
}

You can also access the application with http://localhost:8000/note.

Test in local

There are workflows for each tests (unit test and integration test). These workflows are triggered when you create or update a PR based on some conditions.

Unit test

cd src

pytest -v --cov --cov-report=term-missing unit_tests/

Integration test

docker-compose up -d

bash check-api-endpoints.sh

Frontend

Requirements

node --version
v16.2.0

npm --version
7.13.0

npm install -g @vue/cli
npm install --save axios

cd front
vue add @vue/router
npm install vuex

Deploy on Kubernetes

You can deploy the application on Kubernetes by manual apply or Argo CD

Manual deploy

Deploy MySQL

First, you need to run MySQL container by statefulset.

kubectl create ns database

kubectl apply -k k8s/mysql/overlays/database

kubectl get all -n database

You can login the database test with the following command.

kubectl exec -it mysql-0 -n database -- mysql -uroot -p$(kubectl get secret -n database  mysql-secret -o yaml | grep MYSQL_ROOT_PASSWORD | sed 's/.*.: \(.*\)/\1/' | base64 --decode) test

Deploy the application

After, you successfully deploy MySQL container, then you can deploy the application.

kubectl create ns api-app

kubectl apply -k k8s/fastapi/overlays/api-app/

kubectl get pod -n api-app

kubectl run --restart Never --image curlimages/curl:7.68.0 -it --rm curl sh
If you don't see a command prompt, try pressing enter.
/ $ curl fastapi.api-app.svc.cluster.local:8000/ping
{"ping":"pong!"}/ $ curl fastapi.api-app.svc.cluster.local:8000/notes/1/
{"id":1,"title":"Beyond the legacy code","description":"Awesome book!"}/ $ exit
pod "curl" deleted

Deploy the application by Argo CD

The following instructions may work only when you use minikube as Kubernetes cluster.

Set up Argo CD

kubectl apply -f argocd/setup/namespace.yaml

kubectl apply -n argocd -f https://raw.githubusercontent.com/argoproj/argo-cd/stable/manifests/install.yaml

kubectl port-forward svc/argocd-server -n argocd 8080:443

You can see the console with localhost:8080.

Login Argo CD

Get login password by the following command.

kubectl -n argocd get secret argocd-initial-admin-secret -o jsonpath="{.data.password}" | base64 -d

You can login Argo CD with username admin and the password you got.

Create a project and an application

kubectl create ns database

kubectl apply -f argocd/projects/database/project.yaml

kubectl apply -f argocd/projects/database/mysql.yaml

kubectl create ns api-app

kubectl apply -f argocd/projects/api-app/project.yaml

kubectl apply -f argocd/projects/api-app/fastapi.yaml

Check the status of the applications on UI

Monitoring

You can monitor the application by Prometheuse. The application expose the endpoint (/metrics) for Prometheus. Also, the collected metrics can be shown with Grafana. You can follow the instruction in README in each directory.

How to update the image?

docker login

docker build src/ -t kanata333/fastapi-example:v<version tag>

docker push kanata333/fastapi-example:v<version tag>