Power View supports key performance indicators (KPIs) for tables and matrices. For this scenario, I will create some measures and convert the average duration into a KPI. To make the most of the interaction between charts I will use different perspectives to gain an understanding of the alert behavior. In order to get these perspectives I added configuration data to the SCOM DW model as I did with the performance model.

Overview

A. Add Alert State Dataset to SCOM DW Model

B. Create Columns, Measures and KPI for Alert Duration

C. Create Alert Duration Report View

Step by step

1. Open SCOM DW Model in Excel
2. Click Manage in the Data Model section of the ribbon in the POWERPIVOT menu 

   

3. Click Existing Connections in the Get External Data section of the ribbon in the Home menu 

   

4. Select the SCOM_DW from the PowerPivot Data Connections 

   

5. Click Open
6. Select Write a query that will specify the data to import
7. Click Next >
8. Add the following query

   --Alert State

   SELECT ARS.AlertGuid

          ,ARS.ResolutionState

          ,ARS.TimeInStateSeconds

          ,ARS.TimeFromRaisedSeconds

          ,ARS.StateSetDateTime

          ,ARS.StateSetByUserId

          ,ARS.DWCreatedDateTime

   FROM Alert.vAlertResolutionState ARS

   INNER JOIN Alert.vAlert AL on AL.AlertGuid = ARS.AlertGuid

   WHERE AL.RaisedDateTime > (GETUTCDATE() - 7)
   

9. Enter Alert State in the Friendly Query Name: field
10. Click Finish
11. Click Close when the import completes successfully
12. Click the Diagram View in the View section of the ribbon

    

13. Resize and arrange the datasets
14. Create a relationship by clicking and dragging the field in the From column to the field in the To column 
            

    From	To
    Alert State: AlertGuid	Alerts: AlertGuid

    

1. Click Data View from the View section of the ribbon
2. Add columns to the Alert State dataset and rename them             

   DAX	Column Name
   =SWITCH([ResolutionState],255,"Closed",0,"New",249,"Acknowledged",248,"Assigned to Engineering",247,"Awaiting Evidence",254,"Resolved",250,"Scheduled")	ResState_Name
   =[TimeFromRaisedSeconds]/60	TimeFromRaisedMinutes

3. Change the format for the TimeFromRaisedMinutes to Decimal Number in the Formatting section of the ribbon  

    

4. Click an empty field in the section below the data columns   

   

5. Enter max_dur:=MAX([TimeFromRaisedMinutes]) in the function field (right below the ribbon) to create a measure  

   

6. Add additional measures for other statistics  

   Measure Name and DAX

   avg_dur:=AVERAGE([TimeFromRaisedMinutes])

   min_dur:=MIN([TimeFromRaisedMinutes])

   alert_count_rs:=DISTINCTCOUNT([AlertGuid])

7. Change the format for max_dur, avg_dur, and min_dur to Decimal Number in the Formatting section of the ribbon
8. Right click on avg_dur and select Create KPI…  

   

9. Select Absolute value: under Define target value:
10. Enter 60 in the Absolute value: field  

    

11. Enter 5 in the slider field between the red and yellow ranges and 15 in the slider field between the yellow and green ranges  

    

12. Click OK
13. Select Refresh All from the Refresh pulldown in the ribbon in the Home menu  

    

14. Click Close when the refresh is complete
15. Close the Power Pivot window
16. Click OK in the Power View dialog window  

    

1. Click Power View From the Insert section of the ribbon in the POWER VIEW menu
2. Enter Alerts Closed in Under 30 Minutes in the Click here to add a title field
3. Adjust the text if desired
4. Expand the Alert State dataset 

   

5. Drag the ResState_Name field from the Alert State dataset into the Filters section
6. Select Closed

   

7. Drag the TimeFromRaisedMinutes field from the Alert State dataset into the Filters section
8. Change to the Advanced filter mode, select is less than, enter 30, and click apply filter

   

9. Expand the Entities dataset
10. Drag the TLMEType field from the Entities dataset into the Filters section
11. Select Windows Computer
12. Expand the avg_dur field in the Alert State dataset
13. Select alert_count_rs, avg_duràStatus and avg_dur à Value from the Alert State dataset
14. Expand the Alerts dataset
15. Select AlertName from the Alerts dataset
16. Select FQDN from the Entities dataset
17. Select Matrix from the Table pulldown from the Switch Visualization section of the ribbon 

    

18. Select Rows – Enable Drill Down One Level at a Time from the Show Levels pulldown in the Options section of the ribbon 

    

19. Click on the avg_dur Status column to sort from red to green 

    

20. Click in the open report view space
21. Select alert_count_rs from the Alert State dataset
22. Expand the MP Product dataset
23. Select Product from the MP Product dataset
24. Select Pie from the Other Chart pulldown in the Switch Visualization section of the ribbon 

    

25. Select None from the Title pulldown in the Labels section of the ribbon in the Layout menu 

    

26. Expand the pie chart to half the width of the report view
27. Change the sort order of the pie chart to use alert_count_rs in descending order
28. With the pie chart selected press Ctrl-C then Ctrl-V to cut and paste
29. Click ALL in the Power View Fields section
30. Expand the Server Info dataset
31. Drag the Environment field from the Server Info dataset to the COLOR section
32. Select Remove Field from the pulldown for the Product field in the COLOR section 

    

33. Align the pie charts with room for a column chart underneath
34. Click in the open report view space
35. Select alert_count_rs from the Alert State dataset
36. Expand the Alerts dataset
37. Select Date from the Alerts dataset
38. Select Stacked Column from the Column Chart pulldown in the Switch Visualization section of the ribbon
39. Select None from the Title pulldown in the Labels section of the ribbon in the Layout menu
40. Drag the Date+Hour field from the Alerts dataset below the Date field in the AXIS section
41. Drag the Role field from the Server Info dataset to the LEGEND section
42. Resize the column chart 

    

43. Click on parts of the charts or legends to highlight areas of interest (for example, the Windows alerts) 

    

44. Rename the tab to Short Alerts
45. Save the workbook

This view shows a forecast for CPU utilization for a single server: 

Overview

A. Prepare the Data Model for Power BI

B. Modify a Line Chart in a Report View

C. Configure the Forecast

Step by step

1. Upload the data model (the workbook size for processing has an upper limit of 250MB as of this writing)
2. Go to your Power BI site and navigate to the workbook location 

   

3. Click … then click Enable for the data model 

   

4. Click Enable

   

5. After the report has been enabled click Close  

   

6. Add the data model to Featured Reports if desired 

   

7. Click on the data model to open it
8. Click on the HTML5 version of Power View in the lower right corner 

   

9. Hover over the green arrow in the upper right corner 

   

10. Click Explore the data 

    

1. Select the report view for forecasting from the tabs on the bottom 

   

2. Select the line chart 

   

3. Reduce the number of lines to one by unchecking values or the legend field 

   

1. Click Analysis  

   

2. Select the line chart
3. Click on the right arrow to expand the forecast 

   

4. Change the Confidence Interval to one standard deviation 

   

5. Select different tiles to see the forecast applied to the rest of the view 

   

6. Review the automatic Seasonality value to see the detected data cycle 

   

7. Hover over the exclamation mark in the upper left of the chart to see warnings 

   

I would note a few things about the design decisions for the model:

• The model is generic.  I intended the model as a foundation for customization. 
• The model is meant to scale to large SCOM DWs.  I chose a week to keep it manageable for the >10,000 agent range, but a smaller environment could easily go back a month.  Also, the performance data comes from the daily aggregation, but it could be switched to hourly aggregated data for more detail in smaller environments.
• Workflows are a combination of rules and monitors.  There is some manipulation of RuleRowId and MonitorRowId values to come up with a unique list of workflows.  I used the monitor indicator field to distinguish them (0 = Rule, 1 = Monitor). 

I suggest running the SQL Queries individually before populating the model to see how long they take and how much data is returned. I would suggest adding configuration data at the entity level (see my earlier post), too.

Overview

A. Change Existing Connection

B. Add a Technical Knowledge Dataset for Management Packs

C. Refresh the Data Model and Review Reports

Step by step

1. Start Excel
2. Open the SCOM DW Model file
3. Click Enable Content if there is a security warning 

    

4. Click Manage in the Data Model section of the ribbon in the POWERPIVOT menu 

   

5. Click Existing Connections in the Get External Data section of the ribbon in the Home menu 

    

6. Select the SCOM Data Warehouse from the PowerPivot Data Connections 

   

7. Click Edit
8. Enter an appropriate name in the Friendly connection name: field
9. Enter the SCOM DW server (server\instance if using a named instance) in the Server Name: field
10. Change providers in the Advanced Properties if there is a provider error 

    

11. Select the SCOM DW database (OperationsManagerDW by default) from the pulldown list for the Database name: field
12. Click Save
13. Click Close
14. Select Refresh All from the Refresh pulldown in the ribbon in the Home menu 

    

15. Click Close when the refresh is complete

1. Open a new Excel workbook
2. Copy the ManagementPackDefaultName column from the MPs dataset
3. Paste the ManagementPackDefaultName column in the new workbook
4. Create a column named Product
5. Enter desired values in the cells for any or all management packs (empty cells will show as "Blank" in Power View). For example: 

   ManagementPackDefaultName	Product
   Operations Manager APM Web	SCOM
   Operations Manager APM WCF Library	SCOM
   Operations Manager APM Windows Services	SCOM
   Windows Server 2012 Operating System (Monitoring)	Windows
   Windows Server 2012 R2 Operating System (Monitoring)	Windows
   Windows Server Cluster Disks Monitoring	Windows
   Microsoft SQL Server 2012 AlwaysOn (Discovery)	SQL
   Microsoft SQL Server 2012 AlwaysOn (Monitoring)	SQL
   Microsoft SQL Server 2012 (Monitoring)	SQL

6. Rename the worksheet to MP Product
7. Save the workbook as Tech Knowledge
8. Close the workbook
9. In the Power Pivot model, click From Other Sources in the Get External Data section of the ribbon 

   

10. Select Excel File from the Table Import Wizard then click Next > 

    

11. Click Browse
12. Select Tech Knowledge and click Open
13. Check Use first row as column headers. then click Next > 

    

14. Click Finish then click Close
15. Click on Diagram View from the View section of the ribbon in the Home menu 

    

16. Create a relationship by dragging ManagementPackDefaultName in the MPs dataset to ManagementPackDefaultName in the MP Product dataset 

    

      

1. Select Refresh All from the Refresh pulldown in the ribbon in the Home menu 

   

2. Click Close when the refresh is complete
3. Close the Power Pivot window
4. Click OK in the Power View dialog window 

   

5. Review the reports
6. Click on the Perf Insertions tab
7. Select the column chart 

   

8. Click ALL under the Power View Fields

   

9. Expand the MP Product dataset 

   

10. Drag the Product field to the LEGEND to replace Object

    

11. Analyze the volume of performance data collected 

    

12. Save the workbook

SCOM DW Model and Queries.zip

I use a data model similar to the performance dataset in the Operations Manager Data Warehouse schema (technet.microsoft.com/en-us/library/gg508713.aspx). However, I go from the instances (vPerformanceRuleInstance) to the objects and counters (vPerformanceRule) before going to the data collection rules (vRule). The object\counter pairing forms a unique string for combining with technical knowledge. I should point out that my use of the term "technical knowledge" refers to an arbitrary set of attributes for describing performance counters. For example, I might describe "LogicalDisk\Avg. Disk sec/Transfer" like this:

This helps keep values visible in tiles and legends while communicating the meaning to an audience. The attributes become useful in filtering, too. For this post I will use an Excel workbook for storing technical knowledge, but these could be columns in a SQL table, SQL query or calculated columns in the data model.

Overview

A. Create Performance Data Model

B. Add Measures and Calculated Columns

C. Add Attributes for Object\Counter Values

D. Create Relationships

E. Make a View for All Servers

F. Make a View for Individual Servers

Step by step

1. Set up Excel 2013
2. Create a data source for the SCOM Data Warehouse with the following query:

   /*Perf*/

   SELECT     PERF.ManagedEntityRowId

   ,PERF.PerformanceRuleInstanceRowId

   ,PERF.DateTime

   ,PERF.AverageValue

   ,PERF.MinValue

   ,PERF.MaxValue

   ,PERF.SampleCount

   FROM Perf.vPerfHourly PERF

   INNER JOIN vPerformanceRuleInstance PRI ON PRI.PerformanceRuleInstanceRowId = PERF.PerformanceRuleInstanceRowId

   INNER JOIN vPerformanceRule PR ON PR.RuleRowId = PRI.RuleRowId

   WHERE PERF.DateTime > (GETUTCDATE() - 7)

   ORDER BY PERF.DateTime 

    

3. Add additional datasets for each of the following queries:

   /*Object and Counter*/

   select *

   from vPerformanceRule PR

    

   /*Instance*/

   select *

   from vPerformanceRuleInstance PRI

    

   /*Rule*/

   select RU.RuleRowId

   ,RU.ManagementPackRowId

   ,RU.RuleDefaultName

   from vRule RU

   inner join vPerformanceRule PR on PR.RuleRowId = RU.RuleRowId

    

   /*MP*/

   SELECT distinct MP.ManagementPackRowId

    ,MP.ManagementPackDefaultName

    ,MP.ManagementPackSystemName

   FROM vManagementPack MP

   inner join vRule RU on RU.ManagementPackRowId = MP.ManagementPackRowId

   inner join vPerformanceRule PR on PR.RuleRowId = RU.RuleRowId

    

   /*Entity*/

   select distinct

   ME.ManagedEntityRowId

   ,MET.ManagedEntityTypeDefaultName as METype

   ,TLMET.ManagedEntityTypeDefaultName as TLMEType

   ,ME.TopLevelHostManagedEntityRowId

   ,ME.ManagementGroupRowId

   ,ME.Name

   ,ME.DisplayName

   ,ME.Path

   ,ME.ManagedEntityTypeRowId

   from Perf.vPerfDaily PERF

   inner join vManagedEntity ME on ME.ManagedEntityRowId = PERF.ManagedEntityRowId

   inner join vManagedEntityType MET on MET.ManagedEntityTypeRowId = ME.ManagedEntityTypeRowId

   inner join vManagedEntity TLME on TLME.ManagedEntityRowId = ME.TopLevelHostManagedEntityRowId

   inner join vManagedEntityType TLMET on TLMET.ManagedEntityTypeRowId = TLME.ManagedEntityTypeRowId

   where PERF.DateTime > (GETUTCDATE() - 7)

   order by ME.ManagedEntityTypeRowId

    

Note: The vPerformanceRule and vPerformanceRuleInstance views could be added as tables, but I prefer to use queries in case I want to add conditions later. For example, if I want to exclude LogicalDisk counters or dynamic instances.  

B. Add Measures and Calculated Columns

1. Add a Column to the Perf dataset and rename it

   DAX	Column Name
   =DATEVALUE([DateTime])	Date

2. Click an empty field in the section below the data columns 

   

3. Enter max_max:=MAX([MaxValue]) in the function field (right below the ribbon) to create a measure 

    

4. Add additional measures for other statistics                

   Measure Name and DAX

   avg_max:=AVERAGE([MaxValue])

   avg_avg:=AVERAGE([AverageValue])

   avg_min:=AVERAGE([MinValue])

   min_min:=MIN([MinValue])

5. Add a column to the Object and Counter dataset and rename it

   DAX	Column Name
   =CONCATENATE(CONCATENATE([ObjectName],"\"),[CounterName])	Object+Counter

   
   
6. Add columns to the Entity dataset and rename them

   DAX	Column Name
   =IF([TopLevelHostManagedEntityRowId]=[ManagedEntityRowId],1,0)	TLME
   =IF([TLME]=1,[DisplayName], [Path])	FQDN_Rough
   =MID([FQDN_Rough],1,FIND(";",[FQDN_Rough],1,255)-1)	FQDN
   =MID([FQDN],1,FIND(".",[FQDN],1,255)-1)	Hostname
   =MID([FQDN],FIND(".",[FQDN],1,255)+1,255)	Domain
   =MID([Domain],1,FIND(".",[Domain],1,255)-1)	Domain_Short

     
7. Add a measure to the Entity dataset

   Measure Name and DAX

   Host_Count:=DistinctCount([FQDN])

1. Open a new Excel workbook
2. Copy the Object+Counter column from the Object and Counter dataset
3. Paste the Object+Counter column into the new workbook
4. Create columns for Technology, Category, and Purpose
5. Enter desired values in the columns for any or all counters (empty cells will show as "Blank" in Power View). For example:

   Object+Counter	Technology	Category	Purpose
   LogicalDisk\Avg. Disk sec/Transfer	OS	Storage	Latency
   LogicalDisk\Current Disk Queue Length	OS	Storage	Activity
   ManagementGroupAlerts\AlertsCount	SCOM	MG	Activity
   Memory\Available MBytes	OS	Memory	Capacity
   Memory\PercentMemoryUsed	OS	Memory	Capacity
   Network Adapter\Bytes Total/sec	OS	Network	Activity
   Processor Information\% Processor Time	OS	CPU	Capacity
   System\Processor Queue Length	OS	CPU	Activity
   System\System Up Time	OS	System	Availability

   
   
6. Rename the worksheet to Tech Knowledge
7. Save the workbook as Perf Tech Knowledge
8. Close the workbook
9. In the Power Pivot model, click From Other Sources in the Get External Datasection of the ribbon 

    

10. Select Excel File from the Table Import Wizard then click Next > 

     

11. Click Browse
12. Select Perf Tech Knowledge and click Open
13. Check Use first row as column headers. then click Next > 

     

14. Click Finish then click Close

 D. Create Relationships

1. Click the Diagram View in the View section of the ribbon 

    

2. Resize and arrange the datasets
3. Make the relationships by clicking and dragging the field in the From column to the field in the To column

   From	To
   Perf: ManagedEntityRowId	Entity: ManagedEntityRowId
   Perf: PerformanceRuleInstanceRowId	Instance: PerformanceRuleInstanceRowId
   Instance: RuleRowId	Object and Counter: RuleRowId
   Object and Counter: RuleRowId	Rule: RuleRowId
   Object and Counter: Object+Counter	Tech Knowledge: ObjectCounter
   Rule: ManagementPackRowId	MP: ManagementPackRowId

                                 
4. Review the diagram 

    

5. Save the workbook as Perf 7d Model
6. Close the Power Pivot window

1. From the Insert menu click Power View from the Reports section of the ribbon
2. Select Module from the Themes pulldown in the Themes section of the ribbon 

    

3. Enter Performance Summary in the Click here to add a title field
4. Adjust the text if desired (I used 20 point and bold)
5. Expand the Tech Knowledge dataset 

    

6. Drag the Technology field from the Tech Knowledge dataset into the Filters section
7. Select OS 

    

8. Expand the Object and Counter dataset
9. Drag the Object+Counter field from the Object and Counter dataset into the Filters section
10. Select a representative counter for the LogicalDisk, Memory, Network Adapter, and Processor Information objects 

     

11. Expand the Perf dataset
12. Check the boxes for DateTime, max_max, avg_max, avg_avg, avg_min, and min_min (the order of the values determines their placement in the legend)
13. Select Line from the Other Chart pulldown in the Switch Visualization section of the ribbon 

     

14. Select None from the Title pulldown in the Layout menu 

     

15. Drag Category from the Tech Knowledge dataset into the TILE BYfield for the line chart 

     

16. Resize the tile space and the line chart to make room for a table
17. Click in the open tile space 

     

18. Click ALL in the Power View Fields section 

     

19. Check the box for Object+Counter from the Object and Counter dataset
20. Check the boxes for min_min, avg_min, avg_avg, avg_max, and max_max from the Perf dataset
21. Select None from the Totals pulldown in the Options section of the ribbon 

     

22. Resize the table and arrange it so it doesn't overlap with the line chart 

     

23. Rename the tab to Perf Summary

1. Click Power View From the Insert section of the ribbon
2. Enter CPU Performance by Server in the Click here to add a title field
3. Adjust the text if desired
4. Drag the Technology field from the Tech Knowledge dataset into the Filters section
5. Select OS
6. Drag the Category field from the Tech Knowledge dataset into the Filters section
7. Select CPU
8. Drag the Object+Counter field from the Object and Counter dataset into the Filters section
9. Select Processor Information\% Processor Time
10. Check the boxes for DateTime, avg_max, avg_avg, and avg_min from the Perf dataset
11. Select Line from the Other Chart pulldown in the Switch Visualization section of the ribbon
12. Select None from the Title pulldown in the Layout menu
13. Expand the Entity dataset in the Power View Fields section
14. Drag Hostname from the Entity dataset into the TILE BY field for the line chart
15. Resize the tile space and the line chart to make room for a table
16. Click in the open tile space
17. Click ALL in the Power View Fields section
18. Check the box for FQDN from the Entity dataset
19. Check the box for Object+Counter from the Object and Counter dataset
20. Check the boxes for min_min, avg_min, avg_avg, avg_max, and max_max
21. Select None from the Totals pulldown in the Options section of the ribbon
22. Resize the table and arrange it so it doesn't overlap with the line chart 

     

23. Rename the tab to Server CPU
24. Save the workbook 

The seven day model is the one I built then modified in my previous two posts:

blogs.technet.com/b/drewfs/archive/2014/08/12/scom-performance-data-and-power-view.aspx

blogs.technet.com/b/drewfs/archive/2014/08/14/scom-performance-model-with-configuration-data.aspx

I find long term data provides insights into trends and patterns which may not plainly appear within a week. One might identify behavior tied to a fiscal calendar or seasonal events. Also, the visualizations in Power View may help build a case for capacity changes. Overall, the reports help with data-driven analysis.

I should note that configuration changes become more apparent over time, too. That's why I used to store snapshots of server configuration in an operational database. Sudden changes in performance provoke questions about what happened and why.

Overview

A. Modify SQL Queries in the Performance Data Model

B. Refresh the Data and Review Reports

Step by step

1. Open Perf 7d Model in Excel
2. Save the workbook as Perf 182d Model
3. Select Manage from the Data Model section of the POWERPIVOT menu 

   

4. Select the Perf dataset tab
5. Select Table Properties from the ribbon in the Design menu 

   

6. Enter a new query (Perf.vPerfHourly à Perf.vPerfDaily; 7 à 182) 

   /*Perf*/

   SELECT     PERF.ManagedEntityRowId

   ,PERF.PerformanceRuleInstanceRowId

   ,PERF.DateTime

   ,PERF.AverageValue

   ,PERF.MinValue

   ,PERF.MaxValue

   ,PERF.SampleCount

   FROM Perf.vPerfDaily PERF

   INNER JOIN vPerformanceRuleInstance PRI ON PRI.PerformanceRuleInstanceRowId = PERF.PerformanceRuleInstanceRowId

   INNER JOIN vPerformanceRule PR ON PR.RuleRowId = PRI.RuleRowId

   WHERE PERF.DateTime > (GETUTCDATE() - 182)

   ORDER BY PERF.DateTime

    

7. Click Save
8. Select the Entity dataset tab
9. Select Table Properties from the ribbon in the Design menu
10. Enter a new query (7 à 182)

    /*Entity*/

    select distinct

    ME.ManagedEntityRowId

    ,MET.ManagedEntityTypeDefaultName as METype

    ,TLMET.ManagedEntityTypeDefaultName as TLMEType

    ,ME.TopLevelHostManagedEntityRowId

    ,ME.ManagementGroupRowId

    ,ME.Name

    ,ME.DisplayName

    ,ME.Path

    ,ME.ManagedEntityTypeRowId

    from Perf.vPerfDaily PERF

    inner join vManagedEntity ME on ME.ManagedEntityRowId = PERF.ManagedEntityRowId

    inner join vManagedEntityType MET on MET.ManagedEntityTypeRowId = ME.ManagedEntityTypeRowId

    inner join vManagedEntity TLME on TLME.ManagedEntityRowId = ME.TopLevelHostManagedEntityRowId

    inner join vManagedEntityType TLMET on TLMET.ManagedEntityTypeRowId = TLME.ManagedEntityTypeRowId

    where PERF.DateTime > (GETUTCDATE() - 182)

    order by ME.ManagedEntityTypeRowId

     

11. Click Save

1. Select Refresh All from the Refresh pulldown in the ribbon in the Home menu 

   

2. Click Close when the refresh is complete
3. Close the Power Pivot window
4. Click OK in the Power View dialog window 

   

5. Select Austin from the Themes pulldown in the Themes section of the ribbon of the POWER VIEW menu to distinguish the reports from the seven day model 

   

6. Review the reports to make sure they are still presentable 

   

7. Save the workbook

• Which roles in the production SharePoint environment run low on memory? Which farms and servers within those roles run the lowest?
• What does "DNS\Total Query Received/sec" look like by site and DNS server? How does the data compare to "DNS\Total Response Sent/sec" over the same time period?

One might consider SCOM Groups useful for organizing report information. That works in SSRS reports; however, in Power Pivot and tabular data models, a relationship is a one-to-one or many-to-one lookup. There can't be duplicates in the dataset where one is looking up a value. A server tends to be a member of many SCOM groups. I found it very difficult to use more than one group in a model. Another option within SCOM is the Managed Entity Type or class. The classes don't generally help with business-defined data, though. In any event, a configuration management database (CMDB) in some form will be the source of record. It makes sense to get as close to the source in order to minimize conflicting or outdated information.

I would consider some questions when it comes to configuration data:

• How often does it change?
• Is the change history captured somewhere?
• How are discrepancies (duplicate or missing entries) in the data resolved?

I found it helpful to capture a weekly snapshot of server configurations and store it in a SQL table along with the date of the snapshot as a column. In addition to presenting relevant historical information about servers in a performance report, I could report on things like the growth of the environment, the rate of OS deployments or decommissions, and the ratio of physical to virtual servers. For this post I will use an Excel file, Server Info, with the following example information:                                        

Overview

A. Add a Configuration Dataset to the Performance Data Model

B. Create a Relationship and Refresh Data

C. Make a View for Roles

D. Make a View for Server Statistics

Step by step

1. Open Perf 7d Model in Excel
2. Select Manage from the Data Model section of the POWERPIVOT menu

   

3. In the Power Pivot model, click From Other Sources in the Get External Data section of the ribbon  

   

4. Select Excel File from the Table Import Wizard then click Next > 

   

5. Click Browse
6. Select Server Info and click Open
7. Check Use first row as column headers. then click Next > 

   

8. Click Finish then click Close
    

1. Click the Diagram View in the View section of the ribbon 

   

2. Resize and arrange the datasets
3. Make the relationships by clicking and dragging the field in the From column to the field in the To column 

   From	To
   Entity: Hostname	Server Info: Hostname

   

4. Select Refresh All from the Refresh pulldown in the ribbon 

   

5. Click Close when the refresh is complete
6. Close the Power Pivot window
7. Click OK in the Power View dialog window 

   

8. Save the workbook

1. Click Power View From the Insert section of the ribbon
2. Enter Performance Summary by Role in the Click here to add a title field
3. Adjust the text if desired (I used 20 point and bold)
4. Expand the Tech Knowledge dataset 

   

5. Drag the Technology field from the Tech Knowledge dataset into the Filters section
6. Select OS 

   

7. Drag the Category field from the Tech Knowledge dataset into the Filters section
8. Select CPU, Memory, Network, and Storage
9. Expand the Object and Counter dataset
10. Drag the Object+Counter field from the Object and Counter dataset into the Filters section
11. Select a representative counter for the LogicalDisk, Memory, Network Adapter, and Processor Information objects 

    

12. Expand the Perf dataset
13. Check the boxes for DateTime and avg_avg
14. Select Line from the Other Chart pulldown in the Switch Visualization section of the ribbon 

    

15. Select None from the Title pulldown in the Labels section of the ribbon in the Layout menu 

    

16. Drag Category from the Tech Knowledge dataset into the TILE BY field for the line chart 

    

17. Expand the Server Info dataset
18. Drag Role from the Server Info dataset into the Legend field for the line chart
19. Resize the tile space and the line chart to make room for a table and a bar chart
20. Click in the open tile space 

    

21. Click ALL in the Power View Fields section 

    

22. Check the box for Object+Counter from the Object and Counter dataset
23. Check the boxes for min_min, avg_min, avg_avg, avg_max, and max_max from the Perf dataset
24. Select None from the Totals pulldown in the Options section of the ribbon 

    

25. Resize the table and arrange it so it doesn't overlap with the line chart
26. Click in the open tile space
27. Check the box for Environment from the Server Info dataset
28. Check the box for avg_avg from the Perf dataset
29. Select Stacked Bar from the Bar Chart pulldown in the Switch Visualization section of the ribbon 

    

30. Select None from the Title pulldown in the Labels section of the ribbon in the Layout menu
31. Expand the Entity dataset
32. Drag Hostname from the Entity dataset to the Axis field 

    

33. Change the sort order of the bar chart to use avg_avg in descending order 

    

34. Resize the bar chart and arrange it so it doesn't overlap with the table 

    

35. Rename the tab to Role Summary

1. Click Power View From the Insert section of the ribbon
2. Enter Basic Performance Summary in the Click here to add a title field
3. Adjust the text if desired
4. Drag the Technology field from the Tech Knowledge dataset into the Filters section
5. Select OS
6. Drag the Category field from the Tech Knowledge dataset into the Filters section
7. Select CPU, Memory, Network, and Storage
8. Drag the Object+Counter field from the Object and Counter dataset into the Filters section
9. Select any desired counters 

   

10. Check the box for Counter from the Object and Counter dataset
11. Check the box for FQDN from the Entity dataset
12. Check the boxes for Environment and Role from the Server Info dataset
13. Check the boxes for min_min, avg_min, avg_avg, avg_max, and max_max from the Perf dataset
14. Select Matrix from the Table pulldown in the Switch Visualization section of the ribbon
15. Select None from the Totals pulldown in the Options section of the ribbon
16. Drag Category from the Tech Knowledge dataset into the TILE BY field for the line chart
17. Resize the matrix and tile space and leave room for another matrix above it
18. Select the matrix
19. Press Ctrl-C to copy the matrix
20. Press Ctrl-V to paste the matrix
21. Select Rows – Enable Drill Down One Level at a Time from the Show Levels pulldown from the Options section of the ribbon 

    

22. Drag FQDN to the bottom of the rows list 

    

23. Resize the table so it doesn't overlap with the other table 

    

24. Rename the tab to Basic Summary
25. Save the workbook

Alerts typically carry an expectation that something needs to be fixed.  For this reason, it is useful to customize views for an audience.  I will spend some time using filters to remove distracting data from a view.

This filtered view shows System Center related alerts:

In the interest of clarity I will repeat the data modeling procedure from my earlier posts.  However, I will manipulate the data before creating relationships, because I am going to create a relationship between calculated columns.  For the Power View section I will use SharePoint 2013.     

Overview

A. Add datasets to an Excel 2013 PowerPivot data model

B. Manipulate the data

C. Create relationships

D. Make a Power View report

E. Interact with the report       

Step by step 

A. Add datasets to an Excel 2013 PowerPivot data model

1. Open Excel 2013
2. Click the POWERPIVOT tab
3. Click Manage


 

4. Select From Database then select From SQL Server from the pulldown
   

 

5. Enter a name like SCOM_DW in the Friendly connection name: field
6. Enter the SCOM Data Warehouse server name (with instance and port, if necessary) in the Server name: field
7. Select the name of the SCOM DW (OperationsManagerDW by default) from the Database name: pulldown
8. Click Next
9. On the How to Import the Data dialog select Write a query that will specify the data to import
10. Click Next
11. In the Friendly Query Name field type Alerts
12. Copy and Paste in the following query:    
    --Alerts
    SELECT AL.AlertName
          ,AL.AlertDescription
          ,AL.Severity
          ,AL.Priority
          ,AL.Category
          ,AL.ManagedEntityRowId
          ,AL.MonitorAlertInd
          ,AL.RaisedDateTime
          ,AL.RepeatCount
          ,AL.DWCreatedDateTime
          ,AL.WorkflowRowId
    FROM Alert.vAlert AL
    --WHERE AL.RaisedDateTime > (GETUTCDATE() + 1) -- Dummy condition to return no rows
    WHERE AL.RaisedDateTime > (GETUTCDATE() - 35)

 

13. Click Finish
14. Click Close
15. Click Existing Connections in the Ribbon       
          

 

16. Select SCOM_DW from the PowerPivot Data Connections
            

 

17. Click Open
18. Select Write a query that will specify the data to import and click Next >
19. In the Friendly Query Name type Alert Entities
20. Copy and Paste in the following query:     
    --Alert Entities
    select distinct
    ME.ManagedEntityRowId
    ,ME.Name
    ,ME.DisplayName
    ,ME.Path
    ,ME.ManagedEntityTypeRowId
    from Alert.vAlert AL
    inner join vManagedEntity ME on ME.ManagedEntityRowId = AL.ManagedEntityRowId
    inner join vManagedEntityManagementGroup MEMG on MEMG.ManagedEntityRowId = ME.ManagedEntityRowId
    inner join vManagedEntityType MET on MET.ManagedEntityTypeRowId = ME.ManagedEntityTypeRowId
    order by ME.ManagedEntityTypeRowId

 

21. Click Finish
22. Click Close
23. Repeat steps 15-22 to create datasets Workflows and MPs using these queries:
    --Workflows
    (SELECT RU.RuleDefaultName as WorkflowName
    ,RU.ManagementPackRowId
    ,AL.MonitorAlertInd
    ,AL.WorkflowRowId
    FROM Alert.vAlert AL
    inner join vRule RU on RU.RuleRowId = AL.WorkflowRowId
    where AL.MonitorAlertInd = '0'
    and AL.RaisedDateTime > (GETUTCDATE() - 35))
    union
    (SELECT MON.MonitorDefaultName as WorkflowName
    ,MON.ManagementPackRowId
    ,AL.MonitorAlertInd
    ,AL.WorkflowRowId
    FROM Alert.vAlert AL
    inner join vMonitor MON on MON.MonitorRowId = AL.WorkflowRowId
    where AL.MonitorAlertInd = '1'
    and AL.RaisedDateTime > (GETUTCDATE() - 35))

--MPs
SELECT MP.ManagementPackRowId
  ,MP.ManagementPackDefaultName
  ,MP.ManagementPackSystemName
FROM vManagementPack MP
WHERE MP.ManagementPackRowId in (
  SELECT distinct RU.ManagementPackRowId
  FROM Alert.vAlert AL
  inner join vRule RU on RU.RuleRowId = AL.WorkflowRowId
  WHERE AL.MonitorAlertInd = '0'
  and AL.RaisedDateTime > (GETUTCDATE() - 35)
  )
  OR MP.ManagementPackRowId in (
  SELECT distinct MON.ManagementPackRowId
  FROM Alert.vAlert AL
  inner join vMonitor MON on MON.MonitorRowId = AL.WorkflowRowId
  WHERE AL.MonitorAlertInd = '1'
  and AL.RaisedDateTime > (GETUTCDATE() - 35)
  )

B. Manipulate the data

1. Click Data View in the ribbon
   

 

2. Select the Alerts dataset       
   

 

3. Click an empty field under Add Column
            

 

4. Enter a DAX statement in the function field for the date
   =DATEVALUE([RaisedDateTime]) 

 

5. Press Enter
6. Right click CalculatedColumn1 and select Rename Column
   

 

7. Type RaisedDate and press Enter
8. Repeat steps 3-7 for the following DAX statements and column names
   

DAX	Column Name
=TIME(hour([RaisedDateTime]),0,0)+[RaisedDate]	RaisedDate+Hour
=SWITCH([Severity],2,"Critical",1,"Warning",0,"Information")	Sev_Name
=SWITCH([Priority],2,"High",1,"Medium",0,"Low")	Pri_Name
=[WorkflowRowId]*10+[MonitorAlertInd]	WFID+MAI

 

9. Click an empty field in the section below the data columns
    

 

10. Enter a DAX statement in the function field to create a metric to count the number of alerts
    Alert_Count:=COUNTROWS(Alerts)
    
11. Select the Workflows dataset
    

 

12. Click an empty field under Add Column         
13. Enter a DAX statement in the function field to combine WorkflowRowId with MonitorAlertInd
    =[WorkflowRowId]*10+[MonitorAlertInd]

 

14. Press Enter
15. Right click CalculatedColumn1 and select Rename Column
16. Type WFID+MAI and press Enter
17. Close the PowerPivot window
18. Save the workbook
19. Close the workbook

 

C. Create relationships

1. Click Diagram View in the Ribbon
   

 

2. Click WFID+MAI in the Alerts dataset and drag a line to WFID+MAI in the Workflows dataset      
   

 

3. Click ManagementPackRowId in the Workflows dataset and drag a line to ManagementPackRowID in the MPs dataset     
   

 

4. Click ManagedEntityRowId in the Alerts dataset and drag a line to ManagedEntityRowID in the Alert Entities dataset     
   

 

     
5. Arrange and resize the datasets      
    
   

D. Make a Power View report

1. Click the Create Power View Report icon next to the workbook 
   

 

2. Enter Alerts in the Click here to add a title textbox
3. Expand the Alerts dataset in the Field List then click the checkbox next to Sev_Name
   

 

 

4. Click the Slicer icon in the ribbon       
    

 

5. Click in the open area of the report to deselect any objects      
   

 

6. Click the checkbox next to Pri_Name in the Field List
7. Click the Slicer icon
8. Select the Sev_Name slicer
9. Click Increase Font Size twice
   

 

10. Select the Pri_Name slicer
11. Click Increase Font Size twice
12. Click in the open area of the report to deselect any objects
13. Click the checkboxes next to Alert_Count and Category
14. Click the Pie icon
         

 

15. Change the sort order for the pie chart by clicking on the sort by values
    

 

16. Select the Layout tab
17. Select the Chart Title pulldown then select None
    
18. Click in the open area of the report to deselect any objects
19. Expand the Alert Entities and MPs datasets in the Field List
20. Select Alert_Count from the Alerts dataset and ManagementPackDefaultName from the MPs dataset
21. Click the Matrix icon
     

 

22. Select AlertName and AlertDescription from the Alerts dataset and DisplayName and Path from the Alert Entities dataset
23. Move AlertDescription to the bottom of the ROWS field
    
 

 

24. Drag Sev_Name from the Alerts dataset into the COLUMNS field
    

 

25. Select the Show Levels pulldown then select Enable drill down on rows
           

 

    
26. Select the Home tab
         

 

27. Click on a value in the Total column
     

 

28. Click the Comma Style icon in the Number section of the ribbon
         

 

29. Click the Decrease Decimal icon twice
    

 

    
30. Click the Text Box icon
    

 

31. Enter Alert Summary in the new text box and position it over the matrix
32. Click in the open area of the report to deselect any objects
33. Select Alert_Count and RepeatCount from the Alerts dataset
34. Click the Matrix icon
35. Select AlertName from the Alerts dataset and DisplayName and Path from the Alert Entities dataset
36. Select the Show Levels pulldown then select Enable drill down on rows
37. Select the Home tab
38. Click on a value in the RepeatCount column
39. Click the Comma Style icon in the Number section of the ribbon
40. Click the Decrease Decimal icon twice
41. Repeat steps 39-41 for the Alert_Count column
42. Click the Text Box icon
43. Enter High Repeat Count Alerts in the new text box and position it over the matrix
44. Click the Show Filters icon in the upper right part of the matrix object
    

 

     
45. Expand RepeatCount in the Filters section
     

 

 
46. Click Advanced Filter Mode
    

 

47. Select is greater than or equal to in the top condition pulldown
48. Enter 10 in the value field
49. Click apply filter
50. Click in the open area of the report to deselect any objects
51. Drag RaisedDate from the Alerts dataset into the Filters section
    

 

52. Drag ManagementPackDefaultName from the MPs dataset into the Filters section
53. Select the Styles tab
54. Select Accent 8 from the Themes section      
         

 

55. Select the Background pulldown and click the background on the bottom right
         
56. Arrange and resize the report objects so they don’t overlap
57. Select File –> Save
58. Enter a name without spaces in the File name: field then click Save

Here is my example report:

E. Interact with the report

1. Identify alerts and their sources with Critical severity and High priority
• Click Critical in the Sev_Name slicer     
• Click High in the Pri_Name slicer
• Drill down on the management pack in the alert summary to find the alert name
• Drill down on the alert name to identify the display name of the alert entity
• Drill down on the alert entity to find the path
• Go back to the Management Pack by clicking Drill up three times 
• Undo the selections by clicking Clear Filter for the slicers

 

2. Review alerts by Category
• Click on a slice of the category pie chart or one of the legend items
• Click in the open area of the pie chart object to clear the selection


  

3. Investigate alerts with high repeat counts
• Click the RepeatCount column header to sort by value
• Drill down on the alert name with the highest repeat count
• Drill down on the display name to find the path
• Go back to the alert name by clicking Drill up twice             

 

4. Change the filter settings for the report
• Expand the ManagementPackDefaultName filter
• Select one or more management packs from the List Filter
• Clear the list by clicking the Clear filter icon
• Expand the RaisedDate filter
• Click the Advanced Filter Mode icon
• Change the date ranges with the conditions
• Click apply filter
• Clear the conditions by clicking the Clear filter icon

In another post, “Performance Reporting with Power View” (www.scom2k7.com/performance-reporting-with-power-view/), I used Excel to create report views. This time I switch to SharePoint 2013 for creating views with event data from the SCOM Data Warehouse. The same Power View functions apply, but SharePoint has distinct advantages (rdlx files, Reading Mode, view duplication, export to PowerPoint) which serve larger scale reporting.

Here’s a chronological view of events tiled by management pack:

I’ll take advantage of features added with SQL Server 2012 SP1 (pie charts, drill downs) to simplify analysis. I will make use of the themes to make error counts display in red, too.

By the way, the SCOM DW contains event information (Event Number, Publisher, User Name, Logging Computer, Parameters, Rendered Description) which I haven’t included.  I designed this model for creating high level views to finds trends and tuning opportunities.  I will write about modeling specific technologies and simplifying complex variables in other posts. 

Overview

A. Add datasets to an Excel 2013 PowerPivot data model

B. Create relationships

C. Manipulate the data

D. Upload the Excel workbook to SharePoint 2013 with BI features

E. Make a Power View report

F. Interact with the report     

 
Step by step

A. Add datasets to an Excel 2013 PowerPivot data model

1. Open Excel 2013
2. Click the POWERPIVOT tab
3. Click Manage


 

4. Select From Database then select From SQL Server from the pulldown
   

 

5. Enter a name like SCOM_DW in the Friendly connection name: field
6. Enter the SCOM Data Warehouse server name (with instance and port, if necessary) in the Server name: field
7. Select the name of the SCOM DW (OperationsManagerDW by default) from the Database name: pulldown
8. Click Next
9. On the How to Import the Data dialog select Write a query that will specify the data to import
10. Click Next
11. In the Friendly Query Name field type Events
12. Copy and Paste in the following query:    
    --Events
    SELECT
    EV.DateTime
    ,EV.EventChannelRowId
    ,EV.EventLevelId
    ,ER.RuleRowId
    ,ER.ManagedEntityRowId
    FROM Event.vEvent EV
    INNER JOIN Event.vEventRule ER on ER.EventOriginId = EV.EventOriginId
    WHERE EV.DateTime > (GETUTCDATE() - 65)
    --WHERE EV.DateTime > (GETUTCDATE() + 1) -- Dummy condition to return no rows      
    

    ---

    Note: It is a good idea to test the date range in the query to find what works best in your environment.

    ---

 

13. Click Finish
14. Click Close
15. Click Existing Connections in the Ribbon       
          

 

16. Select SCOM_DW from the PowerPivot Data Connections
            

 

17. Click Open
18. Select Write a query that will specify the data to import and click Next >
19. In the Friendly Query Name type Event Entities
20. Copy and Paste in the following query:
    --Event Entities
    select distinct
    ME.ManagedEntityRowId
    ,ME.ManagementGroupRowId
    ,ME.Name
    ,ME.DisplayName
    ,ME.Path
    ,ME.ManagedEntityTypeRowId
    FROM Event.vEvent EV
    inner join Event.vEventRule ER on ER.EventOriginId = EV.EventOriginId
    inner join vManagedEntity ME on ME.ManagedEntityRowId = ER.ManagedEntityRowId
    inner join vManagedEntityManagementGroup MEMG on MEMG.ManagedEntityRowId = ME.ManagedEntityRowId
    inner join vManagedEntityType MET on MET.ManagedEntityTypeRowId = ME.ManagedEntityTypeRowId

 

21. Click Finish
22. Click Close
23. Repeat steps 15-22 to create datasets Channel, Level, and Rules and MPs using these queries:
    --Channel
    SELECT *        
    FROM vEventChannel

--Level
SELECT *  
FROM vEventLevel

--Rules and MPs
select distinct
ER.RuleRowId
,RU.RuleDefaultName
,MP.ManagementPackDefaultName
,MP.ManagementPackSystemName
from Event.vEventRule ER
inner join vRule RU on RU.RuleRowId = ER.RuleRowId
inner join vManagementPack MP on MP.ManagementPackRowId = RU.ManagementPackRowId  

B. Create relationships

1. Click Diagram View in the Ribbon
   

 

2. Click EventChannelRowId in the Events dataset and drag a line to EventChannelRowId in the Channel dataset
   

 

3. Click EventLevelId in the Events dataset and drag a line to EventLevelId in the Level dataset
   
4. Click ManagedEntityRowId in the Events dataset and drag a line to ManagedEntityRowID in the Event Entities dataset
   
5. Click RuleRowId in the Events dataset and drag a line to RuleRowId in the Rules and MPs dataset
   
6. Arrange and resize the datasets      
          
        
   

C. Manipulate the data

1. Click Data View in the ribbon
   
2. Select the Events dataset
        
3. Click an empty field under Add Column
            
4. Enter a DAX statement in the function field for the date
   =DATEVALUE([DateTime])
          
5. Press Enter
6. Right click CalculatedColumn1 and select Rename Column
   
7. Type Date and press Enter
8. Click an empty field under Add Column
9. Enter a DAX statement in the function field for the date and hour
   =TIME(hour([DateTime]),0,0)+[Date]
10. Right click CalculatedColumn1 and select Rename Column
11. Type Date+Hour and press Enter
12. Click an empty field in the section below the data columns
    
13. Enter a DAX statement in the function field to create a metric to count the number of events
    Event_Count:=COUNTROWS(Events)
14. Close the PowerPivot window
15. Save the workbook
16. Close the workbook

D. Upload the Excel workbook to SharePoint 2013 with BI features

1. Navigate to a SharePoint 2013 PowerPivot Gallery
   
2. Select Files then click Upload Document
    
3. Click Browse…
4. Navigate to the Excel workbook location and select the workbook
5. Click Open
6. Click OK

E. Make a Power View report

1. Click the Create Power View Report icon next to the uploaded workbook
   
2. Enter Event Summary in the Click here to add a title textbox
3. Expand the Level dataset in the Field List then click the checkbox next to EventLevelTitle
   
   
4. Click the Slicer icon in the ribbon
        
5. Click in the open area of the report to deselect any objects
        
6. Expand the Channel dataset then click the checkbox next to EventChannelTitle
   
          
7. Click the Slicer icon
8. Click in the open area of the report to deselect any objects
9. Expand the Events dataset
10. Click the checkboxes next to Date, Date+Hour, and Event_Count
11. Click the Column icon      
    
12. Drag the Date+Hour field from LEGEND to AXIS underneath Date
           
13. Expand the Rules and MPs dataset
14. Drag ManagementPackDefaultName to the LEGEND field
    
15. Click in the open area of the report to deselect any objects
16. Click the checkboxes next to ManagementPackDefaultName and RuleDefaultName from the Rules and MPs dataset
17. Click the checkbox next to Event_Count from the Events dataset
18. Expand the Event Entities dataset
19. Click the checkbox next to Path
20. Click the Matrix icon
    
21. Select the Show Levels pulldown then select Enable drill down on rows
           
22. Click on a value in the Event_Count column
         
23. Select the Home tab
         
24. Click the Comma Style icon in the Number section of the ribbon
         
25. Click the Decrease Decimal icon twice
    
26. Select the matrix
         
27. Press Ctrl-C to copy the matrix
28. Press Ctrl-V to paste the matrix
29. Select the Design tab
30. Click the Pie icon
         
31. Change the sort order for the pie chart by clicking on the sort by values

32. Select the Layout tab
33. Select the Chart Title pulldown then select None
    
34. Select the Legend pulldown then select None
           
35. Select the Styles tab
36. Select Accent 3 from the Themes section
    
37. Select the Background pulldown and click the background on the bottom right
         
38. Arrange and resize the report objects so they don’t overlap

Here is how my example report looks:

F. Interact with the report

1. View the volume of errors and warnings in the system log
• Click Error in the EventLevelTitle slicer
  
• Ctrl-click Warning in the EventLevelTitle slicer
         
• Click System in the EventChannelTitle slicer
         
• Undo the selections by clicking Clear Filter
              

2. Show events by hour for the day with the highest event count
• Double click on the tallest column in the chart
• Go back to the daily view by clicking Drill up
   

3. Change the chart legend
• Select the column chart object
• Drag Path from the Managed Entity dataset to the LEGEND field (this won’t be available in Reading Mode)     
• Click Pop out to zoom in on the chart
  
• Click Pop in to return to the default view
         

4. Drill down to rules then to servers
• Select an MP in the matrix then click Drill down
  
• Double click on the corresponding pie slice
         
• Select a rule in the matrix then click Drill down
           
• Double click on the corresponding pie slice (or entire pie in this case)
         
• Click a pie slice (Ctrl-click to select additional slices) to view events for that server
  

• Go back to the top level of each object by clicking Drill up twice