Cloud Resources (cmdb_ci_cmp_resource) - record management?

[David A2]

We've recently started doing Cloud Discovery for both AWS and Azure and have a significant number of records populating into the Cloud Resource (cmdb_ci_cmp_resource) table, which is a good thing because the cloud teams like the data. My question is, what sort of expectations needs to be set on the management of those CI's? Discovery isn't setting the 'state' field on these, but should I be ensuring the AWS event intake is managing the state on these records?

We're already at 1.5 million records on this table, and the cloud team wants more resource types added. I fear this will be a nightmare to clean up in the future if there isn't some sort of automation helping managed the inventory.

How are other organizations managing this data set? 

[Asi Garty]

As part of cloud inventory discovery, cmdb_ci_cmp_resource has predefined deletion strategy.

By default the strategy is to mark obsolete resource CIs as “absent” (meaning CIs that are no longer discoverable)

This can be changed to “delete” in order for the CIs to be removed from CMDB.
You can change this setting from “sa_ci_to_pattern.list”.

In terms of ongoing maintenance to reserve place, periodically removing unnecessary resources from the discovery inclusion list proven itself well. 
More info here:
https://docs.servicenow.com/bundle/sandiego-it-operations-management/page/product/service-mapping/reference/aws-resource-inventory.html

Keep in mind existing CIs of removed resources will be eventually deleted/marked as absent according to above strategy.

Hope this helps, please feel free to share your findings/additional questions.

[David A2]

Hi Asi, is that "Absent" marking set on the CI anywhere? Both state and install status are blank for these records. How can I track which are absent and which are not?