Controlling Request Rates

Overview

There are various ways to control the requests per second, iterations, and user load in LoadForge.

Naturally, the primary way to control the request rate is with the number of Virtual Users you specify. This guide will focus on additional control beyond that.

Standard wait_time

Typical LoadForge scripts include a wait_time parameter. Examples of this are below:

from locust import HttpUser, task, between, constant

class QuickstartUser(HttpUser):
    wait_time = between(5, 9)
    # or
    wait_time = constant(0.5)
    
    @task(3)
    def index_page(self):
        self.client.get("/")

wait_time is designed to simulate a real browser, where users wait a random amount of time between requests due to them consuming the content on the page.

Zero wait

For the maximum possible request rate (RPS) do not specify a wait time and the script will loop as fast as possible.

You should also consider FastHttpUser if doing that, for a significant improvement to the request rate while sacrificing some of the complex test abilities.

Constant Rate

The constant_total_ips helper will ensure that the total number of requests per second is constant, regardless of the number of users. This is useful when you want to ensure a constant load on your server, regardless of the number of users.

This will not work with the wait_time attribute, as it will override it. Also note, IPS is iterations per second, and means your full test script.

Code Snippet

from locust import HttpUser, task
from locust_plugins import constant_total_ips


class MyUser(HttpUser):
    wait_time = constant_total_ips(5)

    # the below task has two requests, meaning
    # our rps will be around 10 per second (due to 5 IPS above)
    @task
    def t(self):
        self.client.get("/")
        self.client.get("/contact-us")