improvement: IT test should fail early if pods are crashlooping

[apesternikov]

IT tests are timing by timeout if any pods are failing to start. This condition can be detected much earlier by reacting to respective event stream.
This behavior should be configurable by a rule parameter to allow testing of deployments that expect crashlooping as a part of normal workflow.