limits should be parsed and stored by core

currently the limits hooks library parses the userContext of the subnet and the class

this involves 2 aspects that can be avoided:

1. updating the 'local' configuration of the hook each time a subnet is changed

2. using a mutex to protect this 'config update' event

by storing the limits configuration in the subnet and class definition, the parsing and update of the limit is done in a thread safe context (inside a CS either in the command handler or in configure), so no need for a mutex

there is no need to handle configuration changes related to the limis hooks library, and the values can be accessed directly from the server configruation in a thread safe manner