Trouble with Correctness/stale CIs

[Mike Hashemi]

I am working on cleaning up CI correctness and have run into an issue. On the CMDB Health Dashboard, the Correctness widget says, "One or more jobs failed". Clicking the "i" icon, I see "Staleness evaluation status" is "max_failures".

When I select the widget and scroll down, then click the Stale CIs tab, I can see that there are thousands of stale CIs. I selected a class (computer, in this case) and I see CIs listed, that have install status and operational status of "Retired", so I don't understand why they were included in the staleness check.

Is that correct? To reduce the number of stale CIs (which should allow the Correctness job to run), how do I exclude "non-active" (e.g. retired, missing, lost, stolen) states from the staleness calculation?

[Pratiksha]

Hi you need to update the health inclusion rule for staleness : go to cmdb_health_config table, for staleness update the inclusion codition to check only installed ci's. Also check when the stale ci record got created and when the ci got retired. It may possible that someone retired the ci but did not close the task. You can write a simple script to close the stale task where the ci is not in installed state. 

Regards, 

Pratiksha