| || |
Corelation between requests/sec and response time
i'am asking myself the question if i have an application server, with a fixed number of CPUs and a fixed number of connections in the connection pool: "does improving the requests per second metric also means, improving the response time"? This really does matter in case of regression testing, where you would like to have only a single value to see if a patch was worth the work.
Is it also possible to calculate back from a requests per second value to an aproximation of the maximum number of concurrent users an application ca server?
Thanks a lot
Re: Corelation between requests/sec and response time
From my experience requests per second does not necessarily correlate with server response time.
It can have no effect or it could worse the response time if you are over a threshold that creates bottlenecks on the SUT.
It is possible to test what is the maximum number of concurrent users an application by increasing the load to the break point of the system ( benchmark load ).