This is somewhat related to bug 35943. I see this sort of thing regularly:

14:45 <jackmcbarn> why is jenkins-bot failing my commit for stuff i didn't touch?
14:45 <jackmcbarn> https://gerrit.wikimedia.org/r/#/c/97855/
14:46 <ori-l> jackmcbarn: it's the last failure that is causing it to veto
14:46 jackmcbarn looks

The meaning of '(non-voting)' is not very obvious, I think. We ask human reviewers to be explicit in identifying their grounds for -1 patches; I don't see why Jenkins-bot couldn't be held to the same standard :)

IMHO, just exclude non-voting tests from Jenkins's comments. Do people actually look at them?

Version: wmf-deployment
Severity: normal