During testing of ENTMQST-3840 , I have tried scenario with 3 Kafka pods and 3 Zookeeper pods, where user will specify some invalid configuration - f.e. setting request bigger than the limit. When the Zookeper is edited with this invalid configuration, first pod is deleted and after a while, another two pods are deleted, with 1 pod which is trying to be deployed - ending in CrashLoopBackOff state. But, this is not problem just for Zookeeper, but also for all other resources.

Created by Strimzi#7683