Situation with unstable Jenkins runs (on PR)

Hi,

I submitted serveral pull requests lately. And in almost all cases the Jenkins build failed in one way or the other - in all cases unrelated to the PR itself. What is the reason for this and what can be done for testing and accepting PRs in a more reliable way?

looking at the specifics of what happened and helping fixing them on jenkins itself, it that's the problem, or upstream on the packages themselves would be great.

Currently jenkins servers, jenkins configuration (on jenkins.plone.org repo) as well as mr.roboto are all single (un)maintained by me, so any help on that front would be more than appreciated.

I've seen the PR jobs failing, but had no time so far to had a look at them so far.