Skip to main content
Documentation & Lineage · Power BI and Microsoft Fabric

Every model documented.
Every dependency, traced.

Power BI Sentinel automatically documents every semantic model in your estate; DAX measures, Power Query steps, columns, and relationships, then traces the exact path each one takes from source table through to report.

Power BI Sentinel reads metadata, never the data inside your reports. It connects through a Service Principal or SSO using a user token, and writes to a database you own, in your own Azure tenant.

Power BI Sentinel: Documentation Auto-generated
1,842
Measures documented
6,310
Columns tracked
96
Models covered
61
Similar models over 50%
2,417
Unused columns
Recently documented
Finance Model
Mapped
18Measures
14Reports
4Source tables
37Unused cols
Sales Performance
Mapped
27Measures
23Reports
9Source tables
84Unused cols
Customer 360
Mapped
14Measures
11Reports
6Source tables
29Unused cols
Measure dependencies New feature

Every measure,
level by level.

Most measures are not independent, they are built on other measures. Power BI Sentinel walks the whole calculation chain behind any KPI and tells you how far away each link sits, what feeds it directly, what feeds those, and how much of your reporting moves if you change it.

Walks the chain level by level in both directions, direct and indirect
Opens the exact DAX behind any measure in the chain
Counts the visuals and reports that move if you change it, before you change it
Book a product demo
Power BI Sentinel: Measure Dependencies
Showing Net Margin %
4
Upstream
4
Downstream
6
Visuals
3
Reports
Cost of Sales =
SUM ( Costs[Amount] )
Total Cost =
[Cost of Sales] + [Freight Cost]
Feeds into
Net Margin % =
DIVIDE (
  [Total Revenue] - [Total Cost],
  [Total Revenue]
)
Is used by
Margin vs Target =
[Net Margin %] - [Target Margin %]
Margin Health =
IF (
  [Margin vs Target] >= 0,
  "On target",
  "Off target"
)
One measure shown per level · Net Margin % has 8 related measures in total
Power BI Sentinel: Column & Measure Usage
2,417
Columns unused
461
Measures unused
6,310
Columns tracked
1,842
Measures tracked
ObjectReportsStatus
Total Revenue
Measure · Finance Model
14 reports
Used
Net Margin %
Measure · Finance Model
9 reports
Used
Cost Centre Code
Column · Finance Model
0 reports
Unused
Legacy Region Flag
Column · Sales Performance
0 reports
Unused
Scanned across 96 semantic models in shared workspaces
Column & measure usage

See exactly what's used,
and what's dead weight

Every column and measure in every semantic model, mapped against what visuals and reports actually reference it. Find out what's carrying real weight in your estate, and what's just sitting there.

Full column and measure usage documented across every semantic model: DAX measures, Power Query (M) steps, tables, and relationships included
Instantly see which columns and measures are referenced by a live report, and which have no visual pulling from them at all
Built for onboarding, handovers, and audit evidence: a plain, current record of what a model actually does
Book a product demo
Semantic model similarity New feature Discovery tier

See through the chaos
of self-service sprawl.

Self-service publishing leaves you with models nobody planned. Power BI Sentinel scores every semantic model on the tenant against every other one, using the same DAX, columns and measures it already documents, so unnecessary duplication surfaces on its own instead of waiting to be noticed.

Scored on datasources, Power Query, tables, relationships, columns and measures, the same metadata documented everywhere else on this page
Containment scores catch the case where one model sits wholly inside another
Duplicate models mean duplicate refreshes, and refreshes consume capacity you are paying for
Power BI Sentinel: Semantic Model Similarity Mapped
Model A
Finance Model
Finance Reporting
VS
Model B
Finance Model Copy
Exec Reporting
96.40% Overall similarityTwo models, one calculation set, two refresh schedules. High
Where the overlap sits
Datasources100.00%
Power Query100.00%
Tables100.00%
Relationships94.44%
Columns98.65%
Measures88.24%
524 models compared 61 matches over 50%
End-to-end data lineage

Source to report,
the full chain.

Power BI Sentinel builds a live map of every connection in your estate: from raw data sources through gateway connections, refresh schedules, M Code, and SQL queries, all the way to the reports and dashboards built on top.

Maps every dependency from report visual back to source table
Gateway sources, refresh schedules, M Code parsing, and SQL query analysis
See instantly which reports break if a source changes
Shared workspaces only — personal workspaces are not visible to Power BI Sentinel
Power BI Sentinel: Data Source Lineage Scanning
On the roadmap: column-level data lineage traced all the way back to source, extending today's source-table depth down to individual columns.
Definition conflicts

Do you have a single
version of the truth?

Power BI Sentinel compares how the same measure is defined across every semantic model in your estate, and flags where one name hides several different calculations, so "Total Revenue" means the same thing in Finance as it does in Sales.

Cross-model comparison of every documented measure and its underlying DAX
Flags name collisions where a metric is calculated differently model to model
One place to confirm the estate reports a single version of the truth, not several numbers sharing a name
Book a product demo
Power BI Sentinel: Definition Comparison 3 mismatches
Measure "Total Revenue" · in 5 models
Model & workspace Calculation
Finance Model
Finance Reporting
Baseline
CALCULATE ( SUM ( Invoices[NetAmount] ), Invoices[Status] = "Posted" )
Exec Summary
Exec Reporting
Matches
CALCULATE ( SUM ( Invoices[NetAmount] ), Invoices[Status] = "Posted" )
Sales Performance
Sales Analytics
Gross, not net
SUM ( Orders[GrossAmount] )
Customer 360
Customer Insights
Net of credits
SUM ( Invoices[NetAmount] ) - SUM ( Credits[Amount] )
Regional Sales EMEA
EMEA Reporting
No status filter
SUM ( Invoices[NetAmount] )
3 of the 5 models disagree. One measure name, three different numbers: gross instead of net, net of credit notes, and one model counting unposted invoices.
See your estate, fully documented

Stop guessing what
a model actually does

See documentation and lineage mapped across your own estate, live, in a 45-minute demo.

ISO 27001 certified
Cyber Essentials certified
Microsoft SSO & MFA
Your data stays yours