Too many CI jobs started in parallel
I restructured the pipeline (#940 (closed)) and took out unneccessary job dependencies, which is good because it makes the individual pipeline faster by not waiting for jobs that are not needed for other jobs. But it has the disadvantage that other pipelines are not starting because there global limit is reached how many concurrent containers can get started on the runner.
See this pipeline i.e., it was the only one running and is blocking all others from even starting: