I have been developing on a datamart for a couple of months now w/o issues. (Yes, I understand it's Preview.) Performed the tranformations in the datamart and utilized the default semantic model. In Power BI Destkop I consumed the datamart's default semantic model. (DirectQuery to AS-PowerBI Dataset) I then develop the DAX logic and add some local tables with the intent of generating a "golden" dataset; that is, a dataset I will develop one or more reports on.
Everything was working great until this morning. It appears that the datamart lost it's ability to proactively generate a cache model as described in this documentation. I had built an initial report off the golden dataset a few weeks back and it was always respondent. Today, its as if some poured mollases on it. Furthermore, it appears that cache is being built ad hoc as I interact with the report. For instance, clicking on a drop-down slicer will initially take many seconds (depends on the number of records) to populate. Once I have utilzed the slicer and it has populated the domain, then it's back to being performant.
My question is this. How do I determine if a proactive cache of the model has been built and is present in an import model on the service tenent? How do I start to determine where the break is? The documentation states
- Constraints of uniqueness/non-null for particular columns will be enforced in the Import model and will fail the cache building if the data doesn't conform.
I do I determine if my data conformity allows for proactive cache build?
TIA.
