Governing Deployments of Bundles and API deployments

Zhang Zhi
Zhang Zhi Registered Posts: 1

Hi I am using the Dataiku Govern node and I have tried going through the AI Governance Specialist Learning Path. I understand that we are able to govern a deployment of a bundle, however I have not been able to find any information regarding the governing of a API deployment. In addition, looking at the deployed API, there is a status saying that it is not governed, leading me to believe that it is possible to do so.

Additionally, while we are able to govern the deployment of a bundle, how can we create a workflow for modification/updates to that bundle as well as for the retiring/ removal of such deployments? I would like to create a workflow with approvals for that as well

Dataiku version used: 14.2

Dataiku version used: 14.2

Tagged:

Answers

  • Patrick
    Patrick Dataiker, Dataiku DSS Core Designer, Dataiku DSS ML Practitioner, Registered Posts: 1 Dataiker

    Hi,
    Govern is able to secure deployment for bundles as well as API deployment for Saved Models (SM)/Saved Model Versions (SMV) i.e. green diamonds in the Flow. Today, Govern is not able to manage other types of endpoints such as Python endpoints.
    In Govern, for SM/SMV, it works the same way like for bundles. In the Govern node, you can choose to govern them from the "Governable items" page or from the "Model registry".
    You can find more info regarding the type of items you can govern here.

    For the second questions, it would need a bit more details, but with "advanced govern" you can entirely customize your workflows, including with Python hooks and Python actions which would allow to interact with the Deployer (or any other part of the platform) to automated more things.

    Patrick Peinoit — Product Manager

Setup Info
    Tags
      Help me…