Quickstart

Then run the following commands to bootstrap your environment with poetry:

git clone https://github.com/arovira/fastapi-k8s-example-app
cd fastapi-k8s-example-app
poetry install
poetry shell

To run the web application in debug use:

uvicorn app.main:app --reload

Application will be available on localhost in your browser.

Run tests

Tests for this project are defined in the tests/ folder.

This project uses pytest to define tests because it allows you to use the assert keyword with good formatting for failed assertations.

To run all the tests of a project, simply run the pytest command:

$ pytest
================================================= test session starts =================================================
platform darwin -- Python 3.10.8, pytest-7.2.0, pluggy-1.0.0
rootdir: /xxxxx/fastapi-k8s-example-app
plugins: anyio-3.6.2
collected 2 items

src/tests/test_main.py ..                                                                                        [100%]

================================================== 2 passed in 3.94s ==================================================
$

If you want to run a specific test, you can do this with this pytest feature:

$ pytest src/tests/test_main.py

Local deployment with Docker Compose

You must have docker and docker-compose tools installed to work with material in this section. Then just run:

$ docker-compose up --build

Application will be available on localhost in your browser.

Web routes

All routes are available on /docs or /redoc paths with Swagger or ReDoc.

Project structure

Files related to application are in the app or tests directories. Application parts are:

app
├── api
│   └── api.py       - web routes.
│   └── root.py      - web routes.
├── core
│   └── config.py    - application configuration.
│   └── events.py    - start/stop events.
│   └── logging.py   - logging config.
└── main.py          - FastAPI application creation and configuration.
tests
└── test_main.py     - basic pytests

Helm deployment to k8s

In order to deploy to kubernetes, the service image needs to be available on an image repository Since this example uses AWS resources, we will create the ECR repository.

First step is to define some variables we will use related to your AWS account:

export TF_VAR_aws_region=<change to your region>
export TF_VAR_aws_profile=<change to your aws_profile>
export TF_VAR_aws_account_id=<change to your aws_profile>

Execute the following to create it:

cd deploy/terraform/environment
terraform init
terraform apply

Note this uses a terraform module developed on [email protected]:arovira/tfm-aws-ecr-repository.git

Then get the ECR repo name:

export $(terraform output | sed 's/ //g')

Then, go back to the root of the project and build the image (you need docker to do so):

cd ../..
docker build . --tag fastapi-k8s-example-image

Then, authenticate on the ECR repo and push the image:

aws_ecr_base=${TF_VAR_aws_account_id}.dkr.ecr.${TF_VAR_aws_region}.amazonaws.com
aws ecr get-login-password --region $TF_VAR_aws_region --profile ${TF_VAR_aws_profile} | docker login --username AWS --password-stdin ${aws_ecr_base}
docker tag fastapi-k8s-example-image ${aws_ecr_base}/fastapi-k8s-example-app:0.1.0
docker push ${aws_ecr_base}/fastapi-k8s-example-app:latest

Then set your kubernetes context and install via helm:

helm upgrade --install fastapi-example deploy/helm/fastapi-k8s-example-app --set image.repository=${aws_ecr_base}/fastapi-k8s-example-app

If ingress is not enabled, you can access the application with on localhost:8888 after:

kubectl port-forward svc/fastapi-example-fastapi-k8s-example-app 8888:80