Synapse Realtime analytics

Hi Dennes,

Thanks again for helping me think the architecture through. I agree on having all historical data in a lakehouse, realtime data in a KQL dataset. In terms of storage, they indeed shouldn't cross.

However, when reporting, we still would like to have both realtime and historical data in a single dashboard as mentioned in my initial question. As agreed before, requiring a local model is OK.

However, to our experience, this currently significantly degrades the development flow in the following way: If fellow colleagues download a powerbi report from PBI service (in case they would like to add measures etc.), they are prompted to create an additional "local model" (makes sense, dataset is changing). However, this "second" local model creates an unusable dataset when uploading it back to service. (i.e., visuals cannot be displayed, dataset is not accessible, ...).

So the questions still remains: how do we develop a powerbi report (incl. local model) with different developers (not simultaniously), in desktop (for the additional functionality)? 

As I mentioned, simply downloading and overwriting when uploading runs into issues.

I hope we can answer this question to speed up our development process.

Kind regards,
Kjell