Reputation: 243
I have a daily scheduled COGNOS report created by someone else a few months ago. This developer left the company over a month ago. From what I'm told this report ran fine. Then a week before the developer left the company, it began to take longer and longer to run. Now its taking 11-12 hours. There is no documentation for the report that would indicate any possible changes made by the developer that would cause this. And as a very recent new hire; I am, so far, unaware of any db changes that may have caused the issue to begin. I was hoping that someone more experienced could point me in the right direction.
Do I need to investigate changes to the db? Could this be a cache issue? Maybe a permissions problem or just an issue with the created schedule?
Let me know what additional info would be helpful.
Upvotes: 1
Views: 1021
Reputation: 21
Which type of datasource has been used for this report? Multidimensional or relational? Which mode has been used - classic or dynamic? Is this report sending out information to different users via email? Maybe bursting and master detail has been used which is often a performance killer.
Upvotes: 0
Reputation: 544
Start with a simplified version of the report
For example if you have a report with 10 columns
Reduce the report down to 3 columns and compare the performance.
Then add back one data items/column and re-examine If you notice a difference linked to a specific data item, examine the changes in the SQL statement. Look for outer joins or maybe complex expressions with functions which can impact performance
Maybe you are witnessing a logic trap Like a data chasm where you are getting the cartesian product for results because the request is a many to many relationship
It could be the time of day the process is running is now in conflict with another competing process that is taking priority/all of the resources To help test this, try to run the analysis during a non-peak time
Upvotes: 0