OSS Docker Image Repository

[K123AsJ0k1]

To my knowledge OSS-MLOPS-PLATFORM doesn't have own image repository in places like DockerHub, which could be useful for updating custom images like the docker/mlflow image.

Currently, if a user wants to modify and deploy a newer version of MLflow into OSS, they need to first build the docker/mlflow image, load it into the already setup OSS kind cluster and change the deployment yaml at deployment/mlflow/base/mlflow-deployment.yaml using following actions:

cd docker/mlflow
docker build -t mlflow:newest .
kind load docker-image mlflow:newest --name kind-ep
cd /deployment/mlflow/base
nano mlflow-deployment.yaml # Change image version to mlflow:newest
cd ~/oss-mlops-platform
kubectl apply -k deployment

Most of these manual actions could be removed by having a ready made newest docker image of docker/mlflow, which users can then apply along side components like Prometheus and Grafana by simply chancing the image version in their deployment.yamls.