Selenoid is an open source project written in golang and active maintainers are Alexander Andryashin, Ivan Krutov, and  Kirill Merkushev. In the previous article, I tried to explain the usage of docker-selenium and the advantages of it over Selenium GridNow, I will explain another containerization solution to run selenium tests in parallel and show you the advantages of it over docker-selenium

It describes itself as;

“Selenoid is a powerful implementation of Selenium hub using Docker containers to launch browsers.”

I think the most exciting feature of the solenoid is to provide a fresh environment for each test.

When the test starts, the required container is created, the test is executed and the container is removed.

If you were using docker-selenium instead, same containers are used for all suite during the test and this may cause flaky or inconsistent results which are the most dangerous enemy of automation testing I think.


First of all, you need to install docker on your machine. You can follow the official documentation for installationThere are mainly three configurations exist.



You can download cm in non-gui Linux environment with below code snippet:

Note: Chrome images use “/” instead of “/wd/hub” in browser.json

You can run selenoid in a docker container or as a standalone binary.

selenoid tutorial

When we run the start command, cm first checked the existence of docker. It found docker and downloaded required docker images. If docker was not installed, standalone binary would be downloaded.

Our setup is ready but there is no user interface.

Aerokube has another open source project called selenoid-ui. This is a nice user interface to track what’s happening during the test execution. Again, you can use cm to start selenoid-ui.

Alternatively, you can download selenoid-ui binary, give execution permission and execute it.

check http://localhost:8080

If everything goes ok, you should see this:

From that panel, you can see the active sessions, queued tests during execution and if you run your tests successfully across the selenoid, UI panel should look like this:

Now you can stop it by using: ./cm selenoid stop

Other commands can be used with cm:

Additional Configuration

Up to now, we’ve used the most basic features. Now let’s look at what we can do with cm for further configuration. You can list available flags by using ./cm selenoid args

As you can see, we have lots of options to modify it. But most of them cannot be used with cm since they are hardcoded in the configuration manager. So we will use these settings by starting it manually.

Let’s say, you want to start it, but also you want to

  • change the directory of browser configuration file,
  • limit maximum parallel sessions,
  • change the video output directory,
  • limit memory and CPU usage

To make these configuration changes you can start it with the command below:

We’ve simply added desired changes at the end of docker command.

Note: If you are using selenoid binary instead of docker image, you should pass arguments as :

Recommendation on Maximum Parallel Session

Default maximum parallel session limit is 5, but as you know, we can override this value. At this point, the critical question comes to mind. How many tests can I execute in parallel?

Aerokube suggestion for the max parallel session is the number of cores * 1.5 – 2

If you try to exceed this value too much, you will probably be faced a situation like below 🙂

Note: Please keep in mind that this is not only for selenoid. All other solutions have some sort of limitation.

Still, if you want to run more tests in parallel, your options are

  • upgrading your hardware
  • looking for a clustering solution like Ggr which is another Aerokube project.
  • looking for cloud solution like Sauce Labs or Testinium.

But these are not related to this article.

Advance Features

There are some special capabilities that you can disable/enable in your tests. Let’s look at some of them.

Live Browser Screen

We can add visually debug capability with –vnc

./cm selenoid start –vnc

This will download VNC included images and start the selenoid. Don’t forget to set enableVNC capability in your test, which in my case:

ChromeOptions options = new ChromeOptions();

options.setCapability(“enableVNC”, true);

Now chrome sessions have VNC capability. That’s it, easy right?

Note: If you are using a recent version, you may be able to use VNC feature even if you don’t have VNC images. Aerokube team decided that disk consumption of two images is only 10mb and it makes no sense to maintain two types of images. I tried with selenoid/chrome (tag: 67) and it worked.

Custom Test Name

You can set session name by setting capability “name”, which in my case;

ChromeOptions options = new ChromeOptions();

options.setCapability(“name”, testName);

Video Recording

Note: To be able to use this feature, you should have video-recorder image. If you don’t, please use the command below to pull image:

$ docker pull selenoid/video-recorder

Selenoid has a video recorder, you can activate it by passing enableVideo capability as true

ChromeOptions options = new ChromeOptions();


Videos are stored under ~/.aerokube/selenoid/video

video name, frame rate, video screen size and video frame rate are also adjustable.

videoName : string
videoScreenSize : string
videoFrameRate : int

Custom Screen Resolution

screen resolution can be changed with:

screenResolution: <width>x<height>x<colors-depth>

Other than these, there are other fine-tuning options like PerSession Timezone, PerSession Environment variables, Custom DNS Server etc. But I but I will not mention them in this article.


Selenoid is a successful alternative to SeleniumGrid and docker-selenium. It provides a highly configurable execution environment. More importantly, it creates a fresh browser session for each test. Selenoid also has an active and helpful community.  if you got stuck, you can ask for help in their Telegram group.

That’s all I can tell about selenoid. I hope you enjoyed 😉
Kaan Sariveli