Running a Load Test

Running a Load Test

This tutorial describes how to run a simple test.
To learn how to write a custom test, take a look at the custom test tutorial.
Below is a description of the primary fields in the simple test setup.

  • URL – This is the full URL of the page you wish to load test. This should include the URL scheme (e.g. http:// or https://). For example, this might be something like
  • Number of Users – This is the number of users that will be simulated. If you enter 10 here, 10 processes will be started to simulate visitors to the URL provided.
  • Iterations Per User – This is the number of times each user will visit the URL. This is useful in simulating page refreshes. In this case, each user may have cached versions of resources (i.e. images, stylesheets, scripts, etc.) that are found on the page. This is the primary difference between running a 100 user test with a single iteration and a 10 user test with 10 iterations.
  • Ramp Up Time – It evenly distributes the start time of each test over the interval. After that delay, it starts the test, at which point another delay will be introduced based on the min/max delay setting before loading the first URL.
  • Minimum Delay – This sets the minimum delay in milliseconds (1/1000th of a second) that will be simulated per user before each request of the URL.
  • Maximum Delay – This sets the maximum delay in milliseconds that will be simulated per user before each request of the URL.

Advanced Options

There are additional advanced options that are automatically set when doing a simple test.
You have the ability to adjust these settings by clicking on the “Advanced Options” header.

  • Load Assets – This option will parse the HTML response of pages and load the assets on the page. These assets include CSS stylesheets, scripts, and images. This option is enabled by default. A simulated browser cache is used when loading these assets.
  • Use Free Server – If you set 10 or fewer users, 1 iteration, and less than a 30 seconds maximum delay for a simple test, you will be able to run the test on our free server. This allows you to run a test quicker without having to wait for your own server to start up.

Load Agents

RedLine ArchitectureIf you are not running on the free server, you will have the option to define the load agents that will run the test.

Each server that is running a simulated server is called a “Load Agent”. ┬áThese servers run on YOUR Amazon Account – by setting the security access, you give us permission to start and stop the spot instances and keep your costs very low.

Each load agent simulates multiple users simultaneously.

You have the option to add multiple load agents in various geographic locations.

The settings for each load agent are described below.

  • Location – Where the load agent will run.
  • Size – The size of the server. You are only allowed to have one row per Location/Size pair.
  • Number of Servers – Number of servers of this type to run.
  • On-Demand Instance – By default, spot instances are used since they are cheaper, but you can choose to use On-Demand instances. The advantages of this are that the servers will start up quicker and will not be shutdown if your max price is too low.
  • Max Price Per Server – For spot instances, you need to indicate the maximum amount you are willing to pay for a spot instance. If the amount is too low, the test may fail because the bid price is too low or the server may be reclaimed in the middle of a test.
  • Approx. Users Per Servers – This is the approximate number of users to run per server. The sum of all of these should be within X of the total number of users entered in the simple setup portion, where X is the total number of servers. This is approximate since 1 server may be added or subtracted from this amount to meet the overall number of users.