Installation
Internal image repository

Configure an internal image repository

Due to internal IT and security policies, you might need to host the MOSTLY AI container images in an internal repository.

For such cases, you need to edit the values.yaml file in the Helm chart and specify the URL and path to the MOSTLY AI images.

You can find the configuration steps below. Complete before deploying.

Prerequisites

  • Obtain the MOSTLY AI images from your Customer Success Engineer.
  • Work with your IT team to host the images in an internal repository.
  • Make sure that your Kubernetes cluster can access the internal image repository.
  • Obtain the repository URL from your IT team.
  • Obtain the image pull secret for your internal repository from your IT team.

Steps

  1. In the MOSTLY_AI_Helm_charts directory, edit the values.yaml file.
  2. In the repository line, set the internal MOSTLY AI images repository URL.
    repository: <REPOSITORY-URL>
    💡

    For example, repository: mostlyai where mostlyai is the Docker hub registry name.

  3. Tag all MOSTLY AI images with their correct name and version tag in the relevant sections in values.yaml. For example, to tag the mostly-app-enterprise image, you need to have the following:
    APP:
        image:
            name: mostly-app-enterprise
            tag: 113.6.0
  4. Set imagePullSecrets line to match the image pull secrets configured for your repository.
    imagePullSecrets: <INTERNAL-IMAGE-PULL-SECRET>
    💡

    If the repository is different from the MOSTLY AI Nexus repository, the docker_secret is ignored. The imagePullSecrets needs to match the imagePullSecrets of your repository.

    💡

    If you use Docker Hub as an image repository, you can create the secret as follows:

    kubectl create secret docker-registry internal-image-pull-secret \ 
    --docker-server=DOCKER_REGISTRY_SERVER \
    --docker username=DOCKER_USER \
    --docker-password=DOCKER_PASSWORD \
    --docker-email=DOCKER_EMAIL

What's next

Check the deployment checklist for any additional configurations that might be required.

You can then continue with the deployment of MOSTLY AI.