1) You don't need to take a DB backup as during the process of purge and archive all the cases will get archived to a zip folder and purge will happen only once the archive folder gets created. You need to make sure the settings mentioned in below config are properly updated.
2) As purge and archive process have a huge dependency on DB tables, its always suggested performing this activity during the off hours else there might be a chance of occurring performance issues.
3) In case of failure scenarios, enable debug on this key activity Rule_Obj_Activity.ProcessConfig.PegaAccel_Management_Archival.Action to get more information as stated below in logs.
Which class is being processed
which work item is being evaluated
Check if there are any cover items
No of instances of Link-, History
Successfully added to processed list
How long the process ran
Total no of records processed, Failed etc.
In general the failure scenario occurs if a parent case has 2 child case references, but during the purging process if the respective child case entry is not found in DB, then it is considered to be a failure case and the system will not purge that case and proceed with the next case as per the criteria provided.
Incase if there is no clue you are getting from the logs, you can raise an SR to GCS for further investigation.
Thanks for your detailed response. Regarding your response for taking no DB backup required in case of Purge/Archive, could you also provide recommendation when only Purge process is being planned. I understand that if we are running just Purge then the archive to zip folder process is not performed. In such case do you recommend full database backup prior to running Purge?