We have been using the default maximum connections (10) for our connection pool, and I would like to try to increase performance by making this number more realistic.
This is the formula that I have seen: total users * percent of users active at any one time * max number of reports on a dashboard
This seems pretty simple enough, but when it comes to our situation, I am a little confused.
We merge RPD's from several departments into one final deployment RPD. Each of the departmental RPD's have their own
physical data base and connection pool, however, they all point to the same database.
So, for sake of example, here is a simple representation of what is in our final RPD:
Subject Area 1 Business Layer 1 Physical Data Base 1 (Connection Pool 1 pointing to Data Warehouse)
Subject Area 2 Business Layer 2 Physical Data Base 2 (Connection Pool 2 pointing to Data Warehouse)
Subject Area 3 Business Layer 3 Physical Data Base 3 (Connection Pool 3 pointing to Data Warehouse)
There are a total of 2000 OBIEE users: 500 can access Subject Area 1, 1000 can access Subject Area 2, 500 can access Subject Area 3.
When determining the total users for Connection Pool 1, since users from all three connection pools access the same real data base, should I be using Total Users = 2000, or Total users = 500?