Open RobAmey22 opened 5 months ago
Initial investigations point towards Cesium's GPU/hardware acceleration dependency to work efficiently, which is not available in the headless mode (local machine: Linux & Windows & GitLab runners).
@RobAmey22 What led you to believe this could be the cause of the issue here?
Some articles that I read and comments on other issues. @jennifer-shehane
Current behavior
We noticed that our tests in pipeline are taking too long to finish than usual, and jobs getting stuck/failing much frequently.
For instance, a test case which locally takes ~2 minutes (headed run) and ~5 minutes (headless run) is taking ~11 minutes in the pipeline. Initial investigations point towards Cesium's GPU/hardware acceleration dependency to work efficiently, which is not available in the headless mode (local machine: Linux & Windows & GitLab runners).
Otherwise, headless runs used to be much faster than headed every time. Meanwhile, Mac machines are somehow not facing this issue, only Windows and Linux (local machine and GitLab runners).
Pipeline: Linux runners
Desired behavior
Ideally, the pipeline should be much quicker than the local test case execution, no job should get stuck or fail. The pipeline has multiple stages working in parallel on individual runners. Hence, the execution should be quicker and smooth without any stuck cases.
Test code to reproduce
Due to privacy concerns, sharing text code won't be possible.
Cypress Version
12.14.0
Node version
18.16.0
Operating System
Linux
Debug Logs
No response
Other
No response