After running some tests it became apparent that the summary results printed during the test run were not so relevant for my jmeter scripts. In particular they pick up all sorts of samplers, including my data processors and csv loading routines for example. So the numbers reported are not just those hitting my target application. Also, the numbers are very general and I’d like to see how particular transactions are doing, pass and fail rates and response times. This needs a more bespoke solution.
At some point, DB use needs investigating but for now the solution presented on this page works and it’s fairly simple to follow.
As an aside, I will generally be using Splunk and/or Graylog for performance test monitoring and analysis.
I’ll start this topic by showing the final output from this solution, before explaining the settings and code used. This screen shot shows output from Jenkins where I have suppressed the standard summary reports and output specific results for two transactions within the jmeter script. You can see how the test run builds as users come on board and you can see how specific transactions are responding throughout the test: