Forum Discussion
Don't get focused on the BigIP being a GTM or an LTM, or an ASM etc. Think of it more that any BigIP product is also a big3d agent, and it's the agent that gtmd uses to carry out the health check.
The list of big3d agents available is defined as 'all devices in the gtm server list that are type BigIP (or redundant BigIP)'
All monitors that need to be run are allocated to a gtmd processes through a simple hash, and (except for bigip monitors), the gtmd processes allocate the checks to big3d agents, preferring agents that are in the same data centre as the monitor target.
So in regard to your question, if you create a generic host, and assign a gtm http monitor to it, and you have only one GTM in your environment, then the gtmd process on that GTM will assign the monitor to any big3d agent available in the same data centre (which could be itself). If the load on all agents exceeds a threshold, it will also consider big3d agents in other data centres. If a prober pool is assigned, it will limit the available choices to big3d agents on the devices specified in the prober pool.