To handle staled cis in cmdb.

[Roslin Kanushya]

Hi All,

What is the best and correct way to handle/reduce staled cis in cmdb. How can we get list servers with mapped applications like webservice, tomcat servers, tibco cis, mssql db etc. In my instance I am having more than 8k staled cis. So I thought to retire all related cis that are mapped to retired server. While searching in relationship table, I am not getting mapped application. While trying with query builder, Iam getting very few result. Can anyone please assist?

[Mark Bodman]

The stale data usually means that the CI is no longer identified in Discovery, SG Connector or by other means that update the CI regularly from its source.  

I would recommend using the CMDB Data Manager to setup policies and workflow to deal with the Stale CI.  This will allow you to mark it as retired, or move it to archive or deleted when your policy rules have been met.  This overview will help understand how Data Manager works and the types of policies you need to start setting up. 

We also have a new Data Foundations Dashboard to help identify stale CI's and address them.  It's available in the store here.  It supplements the current Out Of Box health dashboard with additional details such as prioritization, and links to KB article playbooks that provide a thorough explanation on how to address each health scan issue, including stale CI's. 

[Roslinkanushya]

Hi Mark,

If I retire (staled ci) tomcat ci , will it be discoverable again? Because, it may not discover for past 60 days due to some discovery error. If that is error is fixed in future and if discovery tries to discover the retired ci, will it get discover?. If yes, I can create a schedule job to retire the staled cis and if it gets discover again, ill change the status to installed again. Will it work?

[Sohail Khilji]

Hi Roslin,

There is a system property that lets you set the stale CI during like 60days or 90days ETC..., its recommended to set the stale CI to ''Retired' instead to delete them. (yes you can do that , they will not more be discovered by Discovery or SGConn)You can also plan for remediation task to CMDB team to review them regularly and see if there is no active relationship on the Stale CI.

Hope i answered efficiently, Kindly Mark correct / helpful if it help....

Sohail Khilji.

[dick linschoten]

although an older post I put my topic here to have it in the same thread.
I watched the youtube movie and tried to apply it on my use case however as in the meantime there is a new datamanager in the workspace it seems to check retirement definitions that checks a specific lifcycle phase before you can run a delete, retire, archive policy. 
as I would like to have different policies to have a stale ci remediate the definitions don't actually work
e.g I would like to clean up discovered servers that remained on initalised status and nobody claimed ownership of them and they haven't been seen by discovery for lets say2 months.
managed by group is empty and most resent discovery relative before 2 months ago and operational status is initialised.
as not all will check on initlised how should we proceed here? as there can only be one defiition per class and that can not contain or statements.