Notebook usage BPA

[natolira]

Semantic-link-labs is great with BPA over semantic models and reports and several other user-friendly functions. Thanks Kovalsky!

What I'm seeing in a large corporation is the fear to adopt Fabric due to the high consumption potential of some bad-written notebooks.

And in many companies regular users don't have admin rights to the workspace. So we need a way to programatically check/update workspace settings to reflect the best practices.

I can think about several ways we can think to enforce the best practices to avoid this high consumption (especially because there are a lot of unexperienced notebook writers that might lead to poor code):

-Set the Default pool for workspace in spark settings to the starter pool single node

-Turn on high concurrency mode to notebooks and data pipelines

-Update the runtime to the most recent GA (this is not updated automatically)

-Check the definition of data pipelines to find multiple notebooks call (especially within loops) and force them to have the same tag to be executed under the same session

-Identify several standard session initialized by the same user or that was timed-out (to notify them to use high concurrency session or remember to manually stop the session when done)

And I'm sure there might be other actions to guarantee wvery workspace respects this guardrails

Thanks again!

[m-kovalsky]

Thanks for your feedback! I believe that pretty much all of these can be checked programmatically via native functions (which call the relevant APIs) within semantic link labs. The one missing is data pipeline definitions (which is coming in the next release).