Creating Policies under the CMDB Data Mananger

Anthony78
Kilo Sage

We have AWS, Azure, VCenter and On-Prem  CIs in ServiceNow. I would like to set up a policy to retire, Archive and delete. Which tables would be best practice to setup these polices . I would like to understand the pros and cons for better understanding.

1 ACCEPTED SOLUTION

Policies set on a larger scale like Hardware [cmdb_ci_hardware] or Application [cmdb_ci_appl] can be ok, especially to introduce the process of CMDB Data Manager; however you want to keep an eye on these as the CMDB scope grows as you may want more nuanced policies over time.

 

Each policy can have a certain complexity to the filter but over time you may decide you need policies at a smaller scope closer to each CI level. It will become a trade-off between # of Policies and # of Classes.

 

For the Virtual Instance, depending on how you are bringing those in, your discovery or integration method may be providing status updates to show that a specific VM is Off or Retired and you may be able to leverage that update to build a relevant Archive policy.

 

To auto-update multiple layers of a CI, you could research Dependent CI management which can cascade-retire or cascade-delete based on the CMDB Data Manager policies. These auto-updates can be processed automatically or be set to generate tasks for a human approval/intervention. (SN Docs: Dependent CIs management (servicenow.com))

View solution in original post

3 REPLIES 3

Marshall Parker
Tera Guru

You will need to identify which CI Class (or multiple classes) you are storing your CIs in from each discovery source.

 

CDMB Data Manager Policies can be setup on a common parent class, but rules should be within the same concept area for management. When you have a different criteria to consider [Class, Source, etc..] you can add a new Policy entry specific to that criteria.

 

Example: If you have On-Prem CIs in both Windows Server & Linux Server, and the criteria for Retire & Archive is the same - you could setup a policy on the parent Server class with the right criteria. But if you have different criteria to consider between those 2 classes, you may want to setup 2 policies directly on each class.

Hi @Marshall Parker ,

So Im thinking to create a policy in  "cmdb_ci_hardware" & cmdb_ci_appl tables so that these tables will be parent for all hardware and software levels .  - Is it okay to create on this table level

But the virtual instances are not coming under this class , how do i retire those ?

 

1.Do i need to create a policy in hardware level and as well in virtual instance level ?

2.do i need to retire servers first or virtual instances first ?

3.Is there a option to retire the ci if either one ci retired 

Policies set on a larger scale like Hardware [cmdb_ci_hardware] or Application [cmdb_ci_appl] can be ok, especially to introduce the process of CMDB Data Manager; however you want to keep an eye on these as the CMDB scope grows as you may want more nuanced policies over time.

 

Each policy can have a certain complexity to the filter but over time you may decide you need policies at a smaller scope closer to each CI level. It will become a trade-off between # of Policies and # of Classes.

 

For the Virtual Instance, depending on how you are bringing those in, your discovery or integration method may be providing status updates to show that a specific VM is Off or Retired and you may be able to leverage that update to build a relevant Archive policy.

 

To auto-update multiple layers of a CI, you could research Dependent CI management which can cascade-retire or cascade-delete based on the CMDB Data Manager policies. These auto-updates can be processed automatically or be set to generate tasks for a human approval/intervention. (SN Docs: Dependent CIs management (servicenow.com))