How to install MLflow on Kubernetes with Helm
Install MLflow in your Kubernetes cluster with Helm, the Kubernetes package manager.
Install and configure these tools first:
Add the repository that holds the MLflow Helm chart:
helm repo add community-charts https://community-charts.github.io/helm-charts
helm repo update
helm search repo mlflow
Create a namespace for MLflow, so it runs apart from your other applications:
kubectl create namespace mlflow
Install MLflow from the "mlflow" Helm chart:
helm install mlflow community-charts/mlflow --namespace mlflow --create-namespace
Check that the MLflow pods are running:
kubectl get pods --namespace mlflow
kubectl get all --namespace mlflow
To change the configuration, copy the chart's values.yaml file, edit the options you need and pass the file to Helm:
helm install mlflow community-charts/mlflow --values values.yaml --namespace mlflow
How you reach MLflow depends on the service type you chose: LoadBalancer, NodePort or Ingress. Ingress needs an Ingress controller running in your cluster.
Follow the instructions in the notes that helm install prints.
We can install MLflow for you. Fully managed, with a 14 day free trial and a 100% money back guarantee
| Extra 1 CPU & 2GB memory | 25 EUR per month |
| Extra 100GB storage (files, database, logs and backups) | 25 EUR per month |
| Private networking, your own IP address or tunnels | 99 EUR per month |
| Private nodes, 1 CPU & 2GB memory (Custom only) | 50 EUR per month |
| Premium SLA: 99.95% uptime, 4-hour P1 response, 14 days of backups | 99 EUR per month |
| Enterprise SLA: 99.99% uptime (multiple replicas), 2-hour P1 response 24/7, 30 days of backups | 499 EUR per month |
Annual Get 2 months for free, 17% discount on a yearly Standard or Premium plan!
Custom and Enterprise: 5% off a 1-year contract, 10% off 2 years or longer, paid per year.
Monthly plans can be cancelled at any time, no commitment.