kubernetes-fastapi
Template for a Python FastAPI with Dockerfile and configuration for Kubernetes
Development setup
To run (in isolation), either:
Run from active Python environment using uvicorn:
pip install -r requirements.txt
uvicorn service.main:app --host 0.0.0.0 --port 8080 --reload
Or build and run the Docker container:
docker build -t 4oh4/kubernetes-fastapi:1.0.0 .
docker run -p 8080:8080 --name kubernetes-fastapi 4oh4/kubernetes-fastapi:1.0.0
Navigate to http://localhost:8080/docs to test the API.
The API responds with a greeting, and the result of a long-running calculation of the largest prime factor of a random integer. You should see a response body similar to:
{
"message1": "Hello, world!",
"message2": "The largest prime factor of 1462370954730 is 398311. Calculation took 0.006 seconds.",
"n": 1462370954730,
"largest_prime_factor": 398311,
"elapsed_time": 0.0057561397552490234
}
Push the container image to Docker Hub
If desired, push the container to Docker Hub yourself, and change all references to the image accordingly. Replace "4oh4" with your Docker Hub ID):
docker push 4oh4/kubernetes-fastapi:1.0.0
You may also need to make the image public as well.
Google Cloud GKE initial setup
Follow the steps in this section if deploying to Google Cloud GKE, or skip if deploying to a ready-configured Kubernetes cluster. From command line, with Google Cloud SDK installed:
gcloud components install kubectl
gcloud config set project my-project-id
gcloud config set compute/zone europe-west2-a
Create a cluster and get credentials for kubectl:
gcloud container clusters create my-cluster-name --num-nodes=3
gcloud container clusters get-credentials my-cluster-name
Kubernetes deployment
kubectl apply -f api.yaml
If working locally, e.g. using minikube, use port forwarding to expose the service:
kubectl port-forward service/kf-api-svc 8080
To scale the deployment, apply a HorizontalPodAutoscaler. Either:
kubectl apply -f autoscale.yaml
or:
kubectl autoscale deployment kf-api --cpu-percent=50 --min=1 --max=10
Load testing
Use locust to simulate a high load on the API
pip install locust
locust
Teardown
kubectl delete deployment kf-api
kubectl delete svc kf-api-svc
kubectl delete hpa kf-api-hpa
Google Cloud clean-up
gcloud container clusters delete my-cluster-name
** Check all resources have been deleted in the console - if in doubt, delete the project as well **
Acknowledgements
Inspiration and code for FastAPI setup: How to continuously deploy a fastAPI to AWS Lambda with AWS SAM.