How flags work
A flag is one check the analyzer runs. A finding is one object that tripped it.
Three unused columns in your model give you one flag — Unused columns — with three findings under it. Expand the flag to see which objects they are.
One object can trip several checks. A measure that is both unused and duplicated appears under two flags, once each.
Severity
Every finding carries a severity.
warning — worth acting on. Unused and duplicated objects, complex DAX, auto date/time tables, heavily loaded pages.
info — context, not a problem to fix. Auto-hidden tables, tables used only in DAX, pages with a moderate visual load, small groups of cards.
Severity is a property of the check, so every finding under one flag has the same severity.
Categories
Flags are grouped by the part of the report they looked at.
| Category | What it covers | Page |
|---|---|---|
| Model | Unused and duplicated tables, columns and measures; auto date/time and hidden tables | Model flags |
| DAX | Measures and calculated columns whose formulas are too complex | DAX flags |
| Visuals | Pages carrying too much, and cards that could be one matrix | Visual and page flags |
| Performance | Auto date/time tables, which inflate model size | Model flags |
Every flag in one table: All flags.
Custom flags
The twelve checks above ship with Soterre PBI Analyzer. You can also write your own - custom flags appear in the results like any other flag, and are left out of the Health Score. See Custom flags.