I have a general question about the analytical model when exporting between environments (say dev to qa or qa to production). It appears that the author of the dashboard/element is added to the permissions of the dashboards, insights, metrics, etc. However when we try and import that model into another environment (promote) we're seeing that fail because the user who authored in a lower environment doesn't exist in the higher level env. Which makes sense to me because dev users aren't the same in dev,qa, etc. Is this something we would have to address ourselves when doing analyticalmodel migrations between environments

Currently there isn’t a different way around it since analytical models are tied to the workspace to which the users belong, therefore, you will need to make sure the users exist across environments.

🎉 New note created.

Thats even more strange in that you have then expect that their IDs are the same between environments as well. We'll most likely strip those out of the model now when bringing across environments. We are already manipulating the model for domain specific values like links and whatnot so this will just be an additional modification.

Personally, I deploy models across environments only with specific logicalModel/analyticsModel APIs, I do not deploy whole workspace(s) definitions. Permissions is one reason, another reason is e.g. to be able to change connected data sources. Generally, with Python SDK you can do anything you want quite easily, but it requires a little of programming. We plan to introduce a new Python SDK package related to LCM (life-cycle management). It should allow users to configure such deployments in config YAML file, no more programming will be required. More requests like this yours help to prioritize it 😉

Interesting and thanks for the insights. Any chance in any language other than python? We're mostly .NET and/or javascript primarily which is why we've been using the 2 model components to simplify the process. Python is hard for us to host in our environments. Programming isn't the issue and we've found quite a few ways to leverage the API's. If the python SDK does some of these items better would be will to revisit the python SDK.

We just introduced the following new feature(Beta): https://gooddataconnect.slack.com/archives/C01M7S4NQAD/p1691682823912929 It is not only about VSCode plugin, we provide CLI as well. It is completely implemented in Typescript. Currently, it covers only LDM and metrics but we plan to extend it with insights/dashboards soon. We would really appreciate your feedback 😉

Been looking at that potentially as we're using VS Code to manage our Snowflake and other integrations already. Potentially could add some additional tools to help us out and do diagnostics.