Is this TPS/throughput calaculation correct?
I want to calculate how many HTTP requests my server can process at a second. I am running the test from 5 clients
So what I did is
1)I summed the column 'HTTP requests' which is under the 'Test summary snapshots'(From all the clients)
2)Then divided the sum by 'total duration of the test'[Elapsed time]
Is this procedure correct to get the HTTP requests processed by the server per second?
If not please advice
Re: Is this TPS/throughput calaculation correct?
No. The value recorded in the snapshot is a running total. In theory, all you need to do is look at the LAST entry in the summary list to see the count of http requests. If running multiple task groups, you will see an entry per task group and they will need to be summed.
After doing this, I cross checked it and it was within 2% of the actual hit count. So, its close but not perfect. To get a more accurate count export http data list and count the number of rows in the file. You may find as I did that the http data indicates a slightly higher hit rate than the snapshot list.