As part of some database benchmarking my team has been trying to do to help 
determine capacity I've been using Jmeter's JDBC Sampler and a Summary Report 
listener to try and gauge at what point we'll begin to see an upward curve in 
what a single node database can handle.

So far our process has gone like this:

1.       Run with ten threads, 10 max connections, for 240 seconds

2.       Record results in excel

3.       Repeat step three until we reach 100 concurrent connections

4.       Plot graph from results


My questions is... is this the "right way" to do this, or will I be getting 
dubious results? Also, is there some way to automate this setup rather than 
changing the numbers and running 10 times?


Thanks,
James

Reply via email to