We are running a load test on MicroStrategy report pointing to Vertica. We are using an open source laod testing tool - LoadUI to test from MSTR web. We are executing same test using JMeter on Vertica to know DB performance for same SQL as of report. But we are experience huge difference between the web report performance and database performance for same SQL. I think we are going wrong in defining same scenarios for both testing (MSTR web and database). Below is the scenario defined in LoadUI for Web report load testing...we would like to know if the scenario we have defined for JMeter is same or different-

LoadUI (to hit MSTR web which is pointing to Vertica):
Base Rate(number of users per min): 200 per min
Total execution time = 10 mins

JMeter (to hit Vertica):
Number of threads (users) = 2000
Ramp up period = 600 seconds
Max number of connections = 2000
Max connection age = 5 seconds

Issue is when we executed these both tests, we observed 17% error rate for MSTR report web testing and 90% error rate for Vertica JMeter test. As per our understanding, db time should not exceed at all than the MSTR time.