[🐛 Bug]: Possible resource leakage using Firefox v. 4.7.0

What happened?

I run two instances of Selenium Grid.

Grid 1: Selenium Grid 3.141.59 with 9 chrome and 9 Firefox nodes (both using tag 3.141.59)
Grid 2: Selenium Grid 4.7.0 with 8 chrome and 8 Firefox nodes (both using tag 4.7.0)

I saw Firefox @ 4.7.0 seems to have some issue with resource consumption & the data looks a lot like leakage.

Have you encountered issues like this?
Might this happen only because some users do not close the driver correctly after testing is done?

Memory consumption

Firefox 3.141.59

Firefox 4.7.0

CPU

Firefox 3.141.59

Firefox 4.7.0

Command used to start Selenium Grid with Docker

I use the Helm files provided.

Relevant log output

See images above

Operating System

Running on Openshift (Kubernetes 1.2X)

Docker Selenium version (tag)

4.7.0

SeleniumHQ/selenium#11270 was fixed with 4.7.1. This could be the resource leak problem.

The client was probably not always closed cleanly: SeleniumHQ/selenium#11345
Why only Firefox is affected is beyond me.

I will try out 4.7.1 and report back.

I tried 4.7.1 and provided it to a subset of the teams. We do not see the same resource leakage here.

See memory usage:

What I also saw during the investigation is that we see these graphs only during times when nodes in 4.7.0 are under heavy load:

(4 GB is the memory limit – it might be that Kubernetes is cleaning up the processes here)

I believe version 4.7.0 is not causing these effects (alone).
I'll provide 4.7.1 to the teams in the next few days and report back if 4.7.1 completely fixed the issue.
If not, I know that some tests do not close the driver as they should.

Updating to 4.7.1 fixed the resource issues!

[github-actions]

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.