I am looking into to moving my multi-threaded python script to locust.
A simple explanation of what my script does is:
When i started looking into locust, I have noticed that the only way to perform each task at its own specific interval, I would need to create a taskset per task.
This brought up an issue of how do i share the auth cookie for the given spawned user between task sets? Since in the long run I also need to share response data between taskset for the given spawned user as it differs between spawned users.
In the sample code below, all of the users spawned by locust, share the same "storage.cookie". Is there a way to keep storage.cookie unique per user, share it with all tasks sets for the given spawned user by locust ? Does locust report on which user is currently executing the task?
from __future__ import print_function
from locust import Locust, TaskSet, task, HttpLocust
import json
def auth(l):
payload = {"username":"some_username","password":"some_password"}
resp = l.client.post('/auth', data = json.dumps(payload))
storage.cookie = # get auth cookie from resp
def do_i_auth(l):
if len(storage.cookie) == 0:
auth(l)
class storage(object):
cookie == ''
class first_call(TaskSet):
def on_start(self):
do_i_auth(self)
@task
def get_api_a(self):
headers = {"Cookie":storage.cookie}
self.client.get('/api_a', headers)
class second_call(TaskSet):
def on_start(self):
do_i_auth(self)
@task
def get_api_b(self):
headers = {"Cookie":storage.cookie}
self.client.get('/api_b', headers)
class api_A(HttpLocust):
task_set = first_call
min_wait = 5000
max_wait = 5000
class api_B(HttpLocust):
task_set = second_call
min_wait = 10000
max_wait = 10000
You can try have your authorization function return the cookie and have it stored in each class separately. Something like this:
from __future__ import print_function
from locust import Locust, TaskSet, task, HttpLocust
import json
def auth(l):
payload = {"username":"some_username","password":"some_password"}
resp = l.client.post('/auth', data = json.dumps(payload))
cookie = # get auth cookie from resp
return cookie
class first_call(TaskSet):
cookie = ""
def on_start(self):
self.cookie = auth(self)
@task
def get_api_a(self):
headers = {"Cookie":self.cookie}
self.client.get('/api_a', headers)
I think the solution here is to not have separate classes for each call, but instead to have the calls as methods on a single class. That way you can store the cookies on the objects (referenced via self.cookie).
This worked for me:
https://gist.github.com/MatrixManAtYrService/1d83abd54adc9d4181f9ebb98b9799f7
I recently implemented cookies in a DotNet application load test script.
Cookies should be passed using dictionary objects.
cookiedict={}
cookiedict['Key1'] = 'Value1'
cookiedict['Key2'] = 'Value2'
#Auth API
self.auth_response = self.gettoken(cookiedict)
self.token = self.auth_response.cookies['RequestVerificationToken']
self.cookies = self.auth_response.cookies
#Login API
cookiedict['RequestVerificationToken'] = self.token
`
self.login_response=self.login(self.user_name,self.password,self.token,cookiedict)
Also note that you need to use HttpSession as well
from locust.clients import HttpSession
self.requests = HttpSession(consumer_cfg.rest_api_url)
executor = self.requests.post
if method == 'PUT':
executor = self.requests.put
elif method == 'GET':
executor = self.requests.get
self._request_proceed(method='GET', url=url, data=formdata,catch_response=catch_response, cookies = CookiesSent,allow_redirects = True)