Hi!
I am taking my time with this because I am highly confused on how our
pipeline worked on some PR without us detecting this. Circle CI
nightly integration tests task has been failing for almost a month
already. But the Jenkins e2e check for PR was fine? I'm very confused
:D Something must be leaking on our pipeline.
We have been accumulating errors for several weeks:
Starting here:
https://circleci.com/gh/syndesisio/syndesis/347407
Full errors here:
https://circleci.com/gh/syndesisio/syndesis/349103
I am working on fixing them, but I am wondering how we have been able
to merge stuff if e2e failed.
Maybe the merge was good because the PR was based on older master
branches with no conflicts, so they passed the tests and we could
merge them?
Which makes me wonder how good those tests are without a pull rebase first.
Or how else have we been able to merge, if nightly e2e with the latest
master fails? Jenkins e2e didn't fail and CircleCI did? Any ideas on
how this could have happened? Or better: how can we prevent this
further?