Skip to content

Latest commit

 

History

History
116 lines (70 loc) · 3.9 KB

install_guide.md

File metadata and controls

116 lines (70 loc) · 3.9 KB

Install Guide

Configuration

The below variables control the configuration of the pruning process. See (Deployment) below for information on when each is needed and created.

These should be defined as AE5 secrets within the service account running the tracking server. Alternatively they can also be set within the anaconda-project.yml project files.

Variables

  1. MLFLOW_TRACKING_URI

    Description

    Remote Tracking Server URI

    Details

    • This should be the static endpoint assigned to the private project deployment.

  2. MLFLOW_REGISTRY_URI

    Description

    Model Registry URI

    Details

    • This should be the static endpoint assigned to the private project deployment.

  3. MLFLOW_TRACKING_TOKEN

    Description

    AE5 Private Endpoint Access Token

    Details

    • Private deployment access token (bearer token)

  4. MLFLOW_TRACKING_ENTITY_TTL

    Description

    The age (measured in days) at which a resource within the MLFlow Tracking Server is considered stale.

    Details

    • Integer value.
    • Measured in days.

    Default

    • 30

Deployment

  1. Use Dedicated Service Account

    • The service account used to run the MLFlow Tracking Server SHOULD also run this service.

  2. Configure Python Environment

    The deployment environment MUST occur within a conda environment with (at least):

     channels:
   - defaults
   - ae5-admin
 dependencies:
   - python=3
   - ae5-tools

  3. Configure AE5 Tools

    Since the project needs to run under the user account created earlier we need to ensure we deploy to this user account. This can be accomplished by either authenticating as the service account itself, or by an AE5 administrator who impersonates the service account at deployment time. See ae5-tools for additional details.

  4. Download Latest Release

    The latest releases can be found here.

  5. [Optional] Slip Stream Customizations

    The default works for most scenarios. However, the archive can be updated and repackaged for specific deployments if needed. This could be useful in scenarios where changes to dependency versions, client specific commands, or default variables must occur prior to deployment.

  6. Upload Project

    This can be accomplished using ae5 tools.

    Example

    ae5 project upload --name "anaconda.mlflow.tracking.prune" anaconda.mlflow.tracking.prune.x.y.z.tar.gz

  7. Create Prune Schedule It is recommended to set up a schedule for the process so that it is occurring regularly.

    • These environment variables MUST be defined as ae5 secrets, within the anaconda-project.yml, or passed to the ae5 job create command are variables (see below).
    Variable
    MLFLOW_TRACKING_URI
    MLFLOW_REGISTRY_URI
    MLFLOW_TRACKING_TOKEN
    MLFLOW_TRACKING_ENTITY_TTL

    Examples

    ae5 job create --command "Prune" --schedule "0 0 * * *" --name "scheduled anaconda.mlflow.tracking.prune" "anaconda.mlflow.tracking.prune"

    ae5 job create --command "Prune" --schedule "0 0 * * *" --name "scheduled anaconda.mlflow.tracking.prune" "anaconda.mlflow.tracking.prune" -variable MLFLOW_TRACKING_ENTITY_TTL=10

Anaconda Project Runtime Commands

These commands are used to start the server and perform the various administrative tasks.

Command Environment Description
Report Runtime Launches Report Only Prune Process
Prune Runtime Launches Prune Process