/mlflow-on-k8s

Deploying MLFlow on Kubernetes

Primary LanguageDockerfile

MLFlow On Kubernetes

High level setup on GCP:

  1. Create Postgres Cloud SQL instance
  2. Fill up k8s/secret.yaml with correct connection string
  3. Setup GCS bucket
  4. Build and push image to GCR
  5. Fill up k8s/deployment.yaml with correct bucket name and image

To build image

docker build . -t mlflow:v1

To run image locally

docker run --name mlflow -d -p 5000:5000 mlflow:v1