About TPS (transactions per second)

The number of transactions processed per second. This value can indicate the number of client requests that the system can process per second under specific load conditions. This is an important indicator to measure the performance of the server, I believe you will often use this indicator during performance testing. But I have always had a question about how this value is calculated. Since the number of transactions per secondAlgorithmNaturally, it is the number of transactions/time ". The number of transactions is well understood. The key is the time. Is the execution time of the entire scenario, or is it just the execution time on the server? As we know, these two times must be different. The former also includes thinktime, pacing time, and network time consumption.
To understand this problem, I checked the help documentation today and saw the above: "The number of transactions per second graph is displayed in every second of the scenario or session step, the number of times each transaction passes, fails, and stops." If you follow this sentence, the answer to the above question should be the latter, that is, in the diagram of transaactions per second, loadRunner samples each time point during the running of the scenario and displays the number of transactions that pass, fail, and stop at this time point.
In addition, I also looked for it in analysis and found that the time display granularity of the chart can also be set. For details, right-click the chart and choose set granularity or press Ctrl + G. I tried to adjust the time granularity to the unit of milliseconds. The result LoadRunner prompts that the display granularity in milliseconds is not supported currently. Therefore, I infer that LoadRunner has transactions per second for this figure, the minimum sampling granularity is 1 second.
If any of the above analyses is incorrect, I hope you will criticize them.