This tutorial provides a walkthrough for installing Gremlin on Windows Server 2019 and performing a Chaos Engineering experiment using a Gremlin CPU attack. The same process can be followed for all supported Windows versions and attack types.
Before you begin this tutorial you need the following:
In this step you'll install Gremlin
gremlin_installer.msi
Gremlin has two forms of authentication for Gremlin clients. Make sure you review the authentication section to figure out which method works for you.
Before you start the gremlin service, you must set some configuration variables. Gremlin recommends using one of the following two methods:
Gremlin's configuration file on Windows is located at C:\ProgramData\Gremlin\Agent\config.yaml
. Please note that ProgramData
is a hidden folder. You can either type in the address on File Explorer or show hidden folders then navigate interactively.
In your installation, you will see an example of this file at C:\ProgramData\Gremlin\Agent\config.yaml.example
copy this file or remove to the .example
suffix, and fill in the file with your desired configuration. See the set of acceptable variables as well as an example on the configuration page.
1# C:\ProgramData\Gremlin\Agent\config.yaml2team_id: 11111111-1111-1111-1111-1111111111113team_certificate: file://C:\ProgramData\Gremlin\Agent\cert.pem4team_private_key: file://C:\ProgramData\Gremlin\Agent\key.pem
Alternatively, if you do not wish to use the configuration file or the init
command, you can set up global system environment variables for the Gremlin service to read. See the full list of environment variables.
Open a command prompt and run gremlin init
. You will be prompted to enter the following values.
Alternatively, you can set up your shell to include these as environment variables to bypass the prompts.
The last step is to restart the Gremlin Service.
Press Ctrl then Esc to display the Start Menu
Start typing services
When Services is selected press Enter or click it
Navigate to the Gremlin Daemon entry.
Right click on it, and click Retart
Alternatively, you can restart the Gremlin Service from the command line (as Administrator):
1net stop gremlind2net start gremlind
The host is ready to be attacked and will be available on your dashboard!
As with attacks on Linux, Windows attacks also have parameters YOU control to give you flexibility in your experimentation, from light to severe impacts. For this article, we will perform a CPU attack. Setting up and executing the other attacks are similar to this.
A CPU attack is designed to consume CPU resources for an amount of time you set and at the level you set. We will use the defaults.
In the Gremlin dashboard (after you login to https://app.gremlin.com), select Create Attack.
Select the Windows host you created and registered as the target for this attack.
In the Choose a Gremlin section, you can select which type of attack you want to run. For this test, select CPU. Adjust the settings are you like.
Notice the red Halt All Attacks button, which stops any running attack and reverts things back to their previous state. This is especially useful if you encounter problems with overly aggressive experiments.
The selection in the above example will consume 75% of the CPU on one core for 60 seconds.
Click on "Unleash Gremlin"
You can watch in the Gremlin UI as the attack runs, and can also see the CPU attack details in the Task Manager on the Windows host.
Congratulations, you just ran your first chaos experiment on Windows!
Feel free to run the experiment many times, varying parameters and seeing how that impacts your instance. Next steps could be to add a full monitoring solution to the instance for better data or perhaps installing the Gremlin CLI and daemon to a different server that you already own and operate (we recommend doing that with a non-production, non-vital server at this stage, should you go this route). You are also welcome to run the other attack types available for Windows, and also set up and run Scenarios on your hosts.
Best wishes and have fun unleashing the Gremlins!
Gremlin empowers you to proactively root out failure before it causes downtime. See how you can harness chaos to build resilient systems by requesting a demo of Gremlin.
Get started