Discovery Table Cleaner for Running Processes leaves orphaned application CIs and relationships behindDescriptionApplication CIs are created automatically by Process Classification, from Running Process records. If an application CI is created this way, the process and classifier are referenced. There is no mechanism to properly deal with the end of life of these records when the process is no longer running. This affects all Discovery customers, and it is normal to see 100k or more application CI's broken in this way in an instance. Several thousand linked to running process records in Absent state that are not deleted yet is also common. When Discovery re-discovers a server, and sees a Running process that it did see last time is no longer running this time, then it sets that cmdb_running_process record to Absent = True. If nothing happens for 30 days, that record is deleted. Due to not cascade deleting, or doing any other special logic to retire process and application, this leaves behind:- Application CIs with broken references to Running Process. The sys_id of the cmdb_running_process record remains in the record, even though it is now deleted.- The Application CI remains in Operational and Installed status, even though it may now not exist on the server. It is never marked Absent.- The 'Runs on' relationship record remains between the application and server Steps to Reproduce 1. Discover a host server running an application, that a process classifier exists for.2. An application CI is created, together with a Runs on relationship with the server.3. Stop the application, and uninstall it.4. Re-discovery the server. The running process record will be set as Absent=true.5. Wait 30 days, or change the age in the table cleaner to delete absent records immediately6. Look at the application record, and nothing will have changed at all, except the running process reference is now broken.WorkaroundThis is fixed as part of Tokyo Release. Related Problem: PRB1418654