Installation¶
Context: The fastapi-traefik application provided by DataScientest
originally used the docker compose format. The application was updated
to be used in a kubernetes cluster environment.
-
Observation:
-
the application uses postgres. The database will be deployed independently to allow for better scaling.
-
The application fastapi uses environmental variables for the DB connection. This meant that we didn’t need to change the code itself, just the environment.
-
Traefik comes preinstalled with k3s, this is what the development environment will use. The staging / production environments will need further testing to find the optimal solution.
The fastapi application has been repackaged to an image and published on a container registry, initially on DockerHub, but eventually it will be published on ECR.
The new image uses the latest python packages and a secure
python base image from Chainguard.
We’re using HELM charts to launch the application in a k8s cluster.
Local development enviroment¶
Preparation of the dev environment¶
Install kubectl¶
TODO
Install helm¶
TODO
Install k3s¶
We install the cluster with etcd and set up the kubeconfig file.
curl -sfL https://get.k3s.io | K3S_TOKEN=SECRET sh -s - server --cluster-init
sudo chmod 644 /etc/rancher/k3s/k3s.yaml
export KUBECONFIG=/etc/rancher/k3s/k3s.yaml
Install cert manager¶
helm install \
cert-manager jetstack/cert-manager \
--namespace cert-manager \
--create-namespace \
--version v1.13.2 \
--set installCRDs=true
Clone the repository¶
Launch the development environment¶
Connect to the local instance¶
Open http://fastapi.localhost:80 in a browser on the local machine (or using a port forward, if it’s on a remote host).
Created: November 17, 2023