BusinessObjects Board

Historical Report Instance Management

How do you manage the number and amount of historical report instances in your Production environments? The management of historical instances can have a significant impact on various performance factors of an environment.

Business Objects (SAP) Support has told me that they have an internal recommendation that we have no more than 50,000 objects in our repository. Since this includes, servers, users, and all other objects in your system, this seems like a very impractically low number. There is no documentation available for this recommendation.

The default value for the number of objects that will be cached for the Central Management Server (CMS) is 100,000. There are documented instructions on how to increase this for better performance SAP KBA 1981894 - “Number of Objects in CMS System Cache” Metric :closed_lock_with_key:

We have close to a million objects in our repository. While this doesn’t appear to have any measurable impact on day to day performance, it does have a significant impact on running the Repository Diagnostic Tool (a few minutes in our non-production environments versus over six hours for our Production environment).

I know we can set limits (we have these set fairly low in our non-production systems), but in our Production system there has been a longer retention period set.

How do other companies manage the historical report instances in their systems? It seems to be something that can get out of hand rather quickly.

Agree that 50,000 seems extremely low, users alone would eat up a big chunk of that limit. I usually use a combination of limits, set most folders to 100 (that keeps daily reports for a little over 3 months), and used APOS to backup and save instances with the limit at 200… jic to give some wiggle room. They also have an option to “archive” the instances which hides them from users so things load faster, but keeps them in the repository.

Im our system , we have a limit of 90 days for instances to be saved.Saving historical instances more than 90 days is a red flag in following best practices in my view. Also, having more historical instances increase the volume of reports and will be challenging during upgrades especially for parallel upgrades.

Srinivas Perumandla

@dtolley What product from APOS do you use with the report instances?
I assume that this is a backup of the instances that would require a restore to view them.

APOS Storage Manager. Yes, you would need to use the admin tool to restore the instance to view it if it was in the online archive or deleted from the CMS. But, it’s selective, you can choose specific instance(s) of the report to restore – you do have to remember that the CMS limits will re-delete the restored instances the next time it runs. Also used it to automatically backup all objects in the entire platform – restoring corrupt/overwritten/deleted content is very easy, saved the day several times! UI is a bit dated. I believe 360, or whatever their new name is, has a similar product.