Hello 
I am running a simulation that has three groups within it. The report is shown below:
I read the docs on these, but I think I am missing something. The response times at the Global Information row are calculated by taking all HTTP requests adding up their response times and dividing with the total number of requests. The same is true for the leaf nodes under the group nodes, but at the specific request level instead.
How is the response time calculated at the group level? The docs say that it is a cumulated response time meaning that if there is a request A that also has 10 inferred resources (I am using inferred resources above) the response time will be from the beginning of request A until the last resource response is received. The docs also mention that pauses are included which confuses me even more as the requests in Scenario #1 have a pause(220, 260) and the mean values are way below!
Would saying that Scenario #1 has a mean response time of 847ms be correct?
I think I completely don’t understand the response times at the group level. Any more detailed clarifications would be much appreciated.
Note that the useGroupDurationMetric is at its default value.
Regards,
George
Cumulated response times is not the sum when it comes to resources fetched in parallel.
It’s the time to fetch the whole resource block.
Hello,
Thanks for the quick response!
Are cumulative values calculated from the start of the first request to the finish of the last in the whole group or just the request along with its inferred resources only (assuming the group contsims more than a single HTTP request)?
group(“group”) {
exec(request1.resources(request1_resources))
.exec(request2.resources(request2_resources))
}
group cumulated response time = request1 + request1_resources (from start to end) + request2 + request2_resources (from start to end)
Perfect!
So pauses of the main requests (i.e., non inferred resources) are not calculated, right?
Thanks again for your help!
Kind regards,
George
Exactly. In short: group cumulated response times = group duration - pauses
Brilliant! Thanks!
Just one suggestion, if this is the correct place to suggest that, maybe the wording:
…cumulated time of a group is the elapsed time from the start of the group’s first request to the end of the group’s last request, including pauses…
in the section, Cumulated Time of https://gatling.io/docs/current/general/timings could change to “not including”?
Thanks again!
Cheers,
George
Indeed, the wording is poor, that’s why I didn’t link the documentation
I’ll fix it ASAP.
Cheers,
Hello again, in case it helps convey the information in the docs in any way this is a note/diagram I have to remember/communicate this.
Regards,
George
