https://blog.mogollon.com.ve/2020/01/10/load-testing-with-python/
Load test tool using aiosonic http client. For drawing charts we use matplotlib and pandas.
Usage of uvloop is highly recommended.
- python>=3.6
pip install aioload
# optional, highly recommended, doesn't work on Windows
pip install uvloopyou need to specify your request in a settings file like config.ini
[http]
sock_read = 30
sock_connect = 3
[test]
# target url for test
url = http://localhost:8080/api/v1/something
# methods: get, post, put, delete
method = post
#
# use body for send body in request
# if body is json, indicate correct header in headers section
# comment body line if you're doing a get request
body = '{"foo": "bar"}'
# query params if needed, this will transform url
# in something like http://localhost:8080/api/v1/something?token=something
[params]
token = something
# headers if needed
[headers]
content-type = application/jsonusage example
> aioload -h
usage: aioload [-h] [-d] [-v] [-n NUMBER_OF_REQUESTS] [-c CONCURRENCY]
[--plot]
testfile
positional arguments:
testfile Test file to be executed
optional arguments:
-h, --help show this help message and exit
-d, --debug true if present
-v, --verbose true if present
-n NUMBER_OF_REQUESTS, --number_of_requests NUMBER_OF_REQUESTS
number of requests to be done, default: 100
-c CONCURRENCY, --concurrency CONCURRENCY
concurrency (requests at the same time), default: 10
--plot draw charts if present
> aioload config.ini -n 3000 -c 100 --plot -v
2019-05-29 17:20:51,662 - __init__:135 - info - 8cf56ded860f41d8a86dab2aed05218f - starting script... -
2019-05-29 17:20:55,301 - __init__:102 - info - 8cf56ded860f41d8a86dab2aed05218f - done - min=14.54ms; max=212.21ms; mean=109.36ms; req/s=600.0; req/q_std=333.7; stdev=24.65; codes.200=3000; concurrency=100; requests=3000;
You can override aioload runner methods, here is an example. Then you should execute the script you made, in this example: python sample/dynamic_test.py conf.ini -v
Python has limits, if your applications is crazy fast like this crystal server, the test will be limited by aiosonic's client speed.
- fork
- create a branch
feature/your_feature - commit - push - pull request
Dependencies are handled with pip-tools
thanks :)