Currently the check is performened only as part of several test cases. It should be executed afterEach. I did a brief look on the code and it seems we need the following changes shown here https://gist.github.com/Frawless/a68a988eb03cb46e5838668e652dcc99 .

The problem is with parallel execution - it will probably fail even the tests, that didn't raise the error in CO log, however, with proper testing we should easily understand which test caused the issue.

This change will for sure need to execute all pipelines for regressions on azure and also on OpenShift.