Request time doesn't include download time when use stream

[zhouweiwei]

Prerequisites

• I am using the latest version of Locust
• I am reporting a bug, not asking a question

Description

I use Locust for performance testing, and one of the cases is downloading files.
However, when using stream=true, the statistics time of Locust does not include the download time of the file.

Command line

locust -f mylocustfile.py --worker --master-host master --host=xxx

Python version

3.8.6

Locust version

2.17

Operating system

mac+docker

Associated issue

#162
067c124

[cyberw]

Hi! When using stream=True this is the expected (intentional) behavior. If that is not what you want, you can use catch_response=True and overwrite the response time using something like this:

with self.client.get(..., stream=True, catch_response=True) as resp:
     start_perf_counter = time.perf_counter()
     # download the response
     download_time = (time.perf_counter() - start_perf_counter) * 1000
     resp.request_meta['response_time'] += download_time

[zhouweiwei]

@cyberw
Hello, download is not work when using catch_response=True and stream=True。
when catch the response , so you can`t read data from respone .
I get 0 size file after the request .

-rw-r--r-- 1 locust locust    0 Nov  3 02:41 DownloadFile_10102_9202_10014_2473.file
-rw-r--r-- 1 locust locust    0 Nov  3 02:41 DownloadFile_10102_9202_10014_3467.file

Do you have any idea about this situation?

[zhouweiwei]

@cyberw
I think there is bug when using catch_response=True and stream=True。
Through the source, it can be seen that request_meta["response_length"] should be response.headers.get("content-length").
But in the execution result, request_meta["response_length"] is 0 and response.headers.get("content-length") is 14031744.

       if kwargs.get("stream", False):
            request_meta["response_length"] = int(response.headers.get("content-length") or 0)
        else:
            request_meta["response_length"] = len(response.content or b"")

with self.client.request("GET", downloadFileUrl, name=name, catch_response=True, context={}, stream=True) as resp, open(outputFile, "wb") as txt:
            
            print(resp.headers.get("content-length")) # 14031744
            print(resp.request_meta['response_length']) # 0

pftest-worker1-1  | 14031744
pftest-worker1-1  | 0

[cyberw]

Yea that looks strange. I guess what that code is actually trying to do is represent the number of bytes actually downloaded. Or something. But I dont know why...

[github-actions]

This issue was closed because it has been stalled for 10 days with no activity. This does not necessarily mean that the issue is bad, but it most likely means that nobody is willing to take the time to fix it. If you have found Locust useful, then consider contributing a fix yourself!