When instantiating an ActiveMQArtemis custom resource instance if you want say 3 brokers and you set size to 3 when you are creating the instance when the 3rd broker pod starts, it will successfully start, then restart, then be usable.

However if you first set size to 1, create and wait for it to be ready, then set size to 2 and wait for it to be ready, and then set size to 3 then neither the 1st, 2nd or 3rd broker pods have an extra restart. In both cases you are left with a normal broker cluster at the point when all three are running and marked ready in OpenShift.

Could be timing in the local OCP environment leading to the readinessProbe timing out or could be a bug related to the scaling state.