Don’t know, the problem looks random to me. Sometimes all jobs pass, sometimes some or even all jobs fail due to network problems (initial git-clone failure, apt-get timeouts etc.). The build history of the project looks better than it really is, I often try the “Restart job” button until it works. In my case, the problematic behavior started on 2021-10-08.
I see. On ruby/ruby repo, we are adding allow_failure for the one Arm64 pipeline. It’s not ideal. But we had to do it, as we see one test failure on the pipeline. So even when the issue sometimes happens on the one Arm64 pipeline, we don’t notice it.
I confirm. Some days it happens more often, some days - less. But the problem is still here, and quite annoying these last two months.
I will do a deep dive, and send this over internally.