When someone triggers PR job, everything runs by default on strimzi-tcn agent, which has smaller resources than the strimzi default agent.

The "tcn" agent is used mainly for jobs where we are running tests on OCP. Otherwise we are using strimzi agent for "deployment" of minikube (which has more memory, so we don't get any OOMs or something like that on the machines).

We should fix this for the PR job - either implementing some trigger which will capture the message and then just runs a different job with correct agent, or set the default agent to be the strimzi one (but we will use bigger machines for the OCP tests, which is not sufficient too).