Need to ignore running processes (PID) when discovering database instances and middleware

[michaelmchugh]

I have a desperate need to disable the process that evaluates running processes (PID numbers) when discovering items such as MS SQL instances, DB2 instances and WebSphere instances.

We have been struggling with this ever since we deployed ServiceNow Discovery several years ago starting with Calgary through Eureka. Unlike hardware devices, no identifiers were available to be defined which would indicate what was already present and what was new. With the introduction of CMDB Identification and Reconciliation when we upgraded to Helsinki, we were hopeful that the new and expanded CI Identification rules would enable us to define these criteria, but unfortunately found that the process that checks the PIDs is still in place, essentially overwriting the rules we have defined.

When we first came up against this issue and brought it to ServiceNow's attention, we were surprised that ServiceNow views these configuration items as "transient". We treat them as if they were similar to hardware with a life cycle all their own and that resources can be attributed to other services and applications and can be charged back against. Tasks such as incidents, requests and changes can be associated to these CI's and they are populated with additional metadata that drive workflows and approvals.

So whenever one of these CI's or thier host servers are recycled either due to normal maintenance or in resolution of an incident or request a new PID is generated and when discovered again a new CI record is created. This has caused us to monitor these on a daily basis and manually remove these newly created CI's before any tasks can be associated against them.

We have looked into modifying the cmdb_running_process table cleaner process without success. We basically just need the evaluation of the PIDs to be either removed or ignored. The new CI Identification rules were our hope, but with the evaluation of PIDs overriding these, they are of little use.

Any ideas or suggestions are welcomes. I would also like to know if others treat database and middle-ware instances in the same manner that we do.

Thanks,

Mike

[csargent]

Hi Mike,

I see this post is fairly old, but I wonder if you developed a solution for this.  We too are monitoring so that duplicate CIs can be cleaned up.  Thanks!

Candy