CMDB Correctness Scorecard - Duplicate denominator

[Larry Youngqui2]

In the CI Class Manager, I have established the CI Identification Rule for independent classes and created an Inclusion Rule (Advanced) that restricts the scope to our list of "managed" classes.

The evaluation of duplicates works as expected.  The "nominator" portion of the percentage is correct.   But the denominator includes more than the total number of CIs in our managed list.   After much searching, we can not determine where or how this number (3109429) is calculated.  Our expectation is approx. 30K. Thus, our percentage is not accurate.

Any ideas on where to validate this number?

FYI: The Completeness Scorecard works exactly as we expect it to.   It's just this one that is not consistent.

Thanks,

Larry

[Mary Vanatta]

Duplication numbers are based on the entire # of CI's in the cmdb_ci table.  Identification rules can be used to create the duplication metrics. 

Note: Be careful when creating Identification inclusion rules. Identification inclusion rules can prevent the identification of certain types of CIs, affecting some features of Discovery and Service Mapping.

Per the documentation site: 

"During duplication detection of independent CIs, the identification engine processes only the CIs that satisfy the identification inclusion rules. For example, you can include only CIs whose operational state is operational. When no identification inclusion rules exist, all CIs are included in the identification process and in the CMDB Health duplicate metric calculations. In the base system, there are no predefined identification inclusion rules. Identification inclusion rules are defined at the class level.

Note: Be careful when creating Identification inclusion rules. Identification inclusion rules can prevent the identification of certain types of CIs, affecting some features of Discovery and Service Mapping."

https://docs.servicenow.com/bundle/london-servicenow-platform/page/product/configuration-management/task/create-id-inclusion-rule.html

Please Mark as the answer and that was helpful.

[Larry Youngqui2]

Thanks, Mary.    I've reviewed the documentation (trust me, many times) and when you say, "Duplication numbers are based on the entire # of CI's in the cmdb_ci table.", that just doesn't seem to be the case for me.

Our CMDB has approx. 5.1m total CI records and the duplication denominator shown in the screen shot shows 3.1m.   I can't account for the 2 million record discrepancy and that's the issue I'm trying to track down.     I want to know how to filter them consistently, accurately, and importantly, have the ability to validate the percentage.

When management asks "How is that percentage calculated?", I can't reply that it's the total number of CI's because it's clearly not.

And I'd agree with you.  When modifying the inclusion rules, it does seem somewhat finicky and browser cache needs clearing, but we're also displaying this same number from multiple workstations.

Larry

[Mary Vanatta]

I also found you have to clear your browser cache, and run the execution jobs several times before the metrics change. 

[Mary Vanatta · Accepted solution]

Larry,

I am thinking that the CI that is counted is marked "Independent" on the Identification Rule. Servicenow Docs states "detection of independent CIs.

If you look at the CI Class Manager> Hierarchy>Hardware>Identification you can see that it is "Independent"

 Whereas,  if you look at the CI "Endpoint" there is no Identification rule and therefore not marked as 
"Independent"

  Now look at the Identification rule for "Storage Device"  that Identification rule shows "Dependent" which would not be counted. 

Navigate to cmdb_identifier_list.do in left nav.  On the Independent tab, sort to True.  The "Hardware Rule" is set to true. So any CI that uses that rule gets counted.