Locust: Strange behavior of "Total Requests per Second" chart

[hoangphucITJP]

Description of issue

Total Requests per Second's RPS drop to below 1 after the test start a while despite the RPS number on top keep rising

Expected behavior

The RPS of Total Requests per Second chart should keep updating according to the RPS number on top

Actual behavior

The RPS of Total Requests per Second chart value does not match the RPS number on top after the test start a while

Environment settings (for bug reports)

• OS: Ubuntu 16.04.3 LTS (GNU/Linux 4.4.0-78-generic x86_64)
• Python version: 2.7.12
• Locust version: 0.9.0

Steps to reproduce (for bug reports)

I don't know what to put here but feel free to ask for any extra information. I can't provide the code though

Screenshot

P/s: Sorry for ignored the template.

[klloveall]

I have also encountered this issue. I dug into the responses and it's happening because the /stats/request endpoint stops returning the "Total" stat block at the end of the stats list. This then causes the graph to show a RPS from a random path instead of from the total like it should. Interestingly, as is shown in the screenshot from @hoangphucITJP, the RPS in the header is still correct.

I tried to dig into the Python code and figure out what's causing it there, but haven't had much success there yet (mostly b/c I'm only running this in AWS, not locally so my debugging options are somewhat limited).

Please let me know if there's any suggestions as to how to solve this or any other debugging information I can provide.

[supermasita]

@cgoldberg I think my simple change should work as a fix (it has solved the issue for me). Thanks.

[supermasita]

@hoangphucITJP @klloveall - let me know if manually applying my change fixes your issue. I think my PR is correct, but it is not building for unrelated reasons. Thanks.

[sweetib]

we are also facing this issue, when is the PR supposed to be merged?

[victor61zhong]

any update for this issue ?

[canstudios-simonm]

I also see the RPS drop off as shown in the graph every time we run Locust. It makes tracking the average RPS value over a run very hard without it working.

[cyberw]

I cant reproduce this (on MacOS, using locust 0.12.1). Is this only an issue on Linux? Or is it something related to the test plan?

[heyman]

I believe this could have the same cause to #1060 which has now been fixed.