I choose duration policy as " By Time". Now If my population stop policy is immediate, then after time finishes, I think my all running tranctions will be stopped immediately and goes to end. That is understood. Now those transactions which were partially finished and stopped abruptly, will they counted in my final average response time per transaction data ? If they are counted then they are actually not finished and will have lesser response time. If you won't count it then it had already some time spent in running and that effort lost and we have lesser transaction count. So poorer result. What is Neoload duing actually?
Transactions that are not finished are excluded from the results. Your test must be long enough to execute enough full transactions and ignore the last ones.
christophe answered
Alpesh posted a new question