While using the Vertica ADO.NET client driver with connection pooling enabled (MinPoolSize of 20 & MaxPoolSize of 100) it was taking ~15 seconds to get results while running 30 queries in parallel. With connection pooling disabled, I was getting results in ~5 seconds for the same 30 queries when called parallelly.
On checking Vertica's query history I could see that the queries themselves are getting completed in <100 ms.
My expectation was that connection pooling should improve the overall performance since the connections don't need to be dropped & recreated. Any reasoning for this degraded performance we are facing when the pooling is on?
We are using Vertica.Data version 9.2.3
