Starting with z/OS® V1R12
WLM considers the resource group maximum and whether the projected
increase in service demand will not exceed the resource group maximum.
When the service is already capped due to the resource group maximum,
then no additional initiators are started.
Note: This resource maximum check is not done on a LPAR for a service
class if WLM has not started an initiator for that particular service
class on that LPAR, or if WLM has only started initiators in that
service class on other LPARs in the sysplex. In this case, WLM can
start an initiator for that service class on that LPAR even if resource
group capping is already active for that service class. Otherwise,
WLM cannot determine how much capacity is being used on average by
a batch job for that service class.