Edit on GitHub

K8s Installation using Helm


DVC Studio Images

Access to the DVC Studio Docker images need to be provided by the Iterative team to enable the installation.


You can deploy DVC Studio Self-hosted on any Kubernetes cluster (version => 1.20) that has an ingress controller configured and installed.

The choice of ingress controller is up to you. We recommend the ingress-nginx controller, but DVC Studio will work on other ingress controllers as well.

You'll also need to have the following CLI tools configured and installed:

  • Helm
  • Kubectl


Create a DNS record pointing to the external IP address of your ingress controller. This hostname will be used for DVC Studio.

1. Create a namespace

We'll install DVC Studio and related components in a dedicated studio namespace. Let's create it now:

$ kubectl create namespace studio

If you want to install DVC Studio in any other namespace, modify the --namespace flag in the following commands accordingly

2. Create a Docker registry secret

Configure Docker credentials for pulling images from our private registry:

$ kubectl create secret docker-registry iterativeai \
    --namespace studio \
    --docker-server=docker.iterative.ai \
    --docker-username=<username> \

Replace <username> and <password> with the credentials you've received from the Iterative team.

3. Add the Iterative Helm repository

$ helm repo add iterative https://helm.iterative.ai

3. Install the DVC Studio Helm chart

In this example, we'll configure DVC Studio to integrate with your Git Forge of choice and use the built-in Postgres, Redis, and Minio.

For a comprehensive list on DVC Studio's configuration options, please see the Configuration documentation.

Let's start by creating a YAML configuration file, values.yaml, with the following contents:

  - name: iterativeai

  host: '<DVC Studio hostname>'
    SELF_HOSTED_LICENSE_KEY: '<License key provided by the Iterative team>'

Here, the SELF_HOSTED_LICENSE_KEY is needed for Studio to correctly recognize your license and allow you to create teams in Studio with a number of available seats equivalent to your purchased seat limit.

Next, you need to integrate our Git Forge with your Studio instance. Depending on the Git Forge you are using follow the corresponding instructions in the Git Forges Configuration documentation.

Finally, deploy DVC Studio with the command:

$ helm install studio iterative/studio \
    --namespace studio \
    -f values.yaml

๐ŸŽ‰ You can now access DVC Studio in your browser on the domain (hostname) that you've configured.


๐Ÿ› Found an issue? Let us know! Or fix it:

Edit on GitHub

โ“ Have a question? Join our chat, we will help you:

Discord Chat