FastAPI web service that is deployed with Kubernetes (K8s) on the local Minikube cluster.
The FastAPI web service has APIs to log information. The log file(s) will be rotated later using a scheduled K8s Job. Alternatively, you can use Python logging handlers such as RotatingFileHandler that support the rotation of log files.
The deployment contains local persistent volume (PV) with ReadWriteMany access mode. The PV resides in the Minikube cluster under the /persistent_volume/logs path. The persistent volume claim (PVC) requests storage from the local PV. The deployment has one pod replica and pulls the docker image from the Minikube cluster. Hence the image pull policy is set to: imagePullPolicy: Never. It mounts the /mnt/fastapi_app logging folder to the PVC. Then, the pod is exposed as service, listening on the 8000 port.
A scheduled job (CronJob) performs log rotation on the logging folder, and it executes every minute. The job also mounts the logging directory and a log rotation config file (ConfigMap). The configuration file is put under the /etc/logrotate.d folder. The copytruncate option is important because it can be used when some program cannot be told to close its log file and thus might continue writing (appending) to the previous log file forever. In other words, the web service can't write to the log file if the option is omitted.
There is an additional archiver job that compresses the log folder content and then deletes them. The archiving is implemented in Python and containerized to be deployed as a cron job. It stores the archived folders in the archives directory using the archived_<current_date>.tar.gz file format. The archive folder is also attached to the mounted path.
- Docker: Install the Docker engine: https://docs.docker.com/engine/install/
- Kubernetes: Download the Kubernetes command line tool: https://kubernetes.io/releases/download/
- Minikube: Install the local Minikube cluster: https://minikube.sigs.k8s.io/docs/start/
Create a virtual environment:
python -m venv .venv
Activate the virtual environment:
source .venv/bin/activate # Linux
source .venv/Scripts/activate # Windows
Install the dependencies:
pip install -r app/requirements.txt
Serve the FastAPI web service using Uvicorn web server:
uvicorn app.main:app --reload
Go to the generate log folders API and create log folders. After the generation, archive the log folders:
python archiver/main.py
Verify the content of the compressed file:
tar -tzf archives/<archived_[CURRENT_DATE].tar.gz>
Build the FastAPI docker image:
docker build -t fastapi -f app/Dockerfile .
Create a container from the docker image:
docker run -p 8000:8000 fastapi
Visit the http://localhost:8000/docs SwaggerUI website. Generate logs via APIs.
Build the archiver docker image:
docker build -t archiver -f archiver/Dockerfile .
Create a container from the docker image:
docker run -it archiver sh
Initiate the Minikube cluster:
minikube start
Reuse the Minikube's Docker daemon:
eval $(minikube docker-env)
docker build -t fastapi-image -f app/Dockerfile .
docker build -t archiver-image -f archiver/Dockerfile .
Go inside the cluster:
minikube ssh
Create the persistent volume folder inside the Minikube cluster:
sudo mkdir -p /persistent_volume/logs
Create the K8s resource:
kubectl apply -f deployment.yml
Tunnel the service through Minikube to access it from the browser:
minikube service fastapi-service
Open the SwaggerUI and generate logs.
kubectl delete -f deployment.yml # delete the k8s workloads
kubectl get all # listing k8s workloads
kubectl rollout restart deployment fastapi-deployment # restart deployment
kubectl describe pod <pod_name> # describe pod info
kubectl logs <pod_name> -p # get logs from pod
kubectl exec -it <pod_name> -- sh # attach shell to pod
docker rmi -f <image_name> # remove image forcefully
minikube image ls # list minikube images
minikube ssh # step into minikube
minikube stop # stop minikube
minikube delete # delete the VM, and associated files
rm -rf $HOME/.minikube # purge minikube
eval $(minikube docker-env -u) # deactivate docker daemon reuse