core hours reporting is incorrect in endpoint reports

[benclifford]

Describe the bug
When a manager shuts down, the two core hours fields have the core hours from those managers removed.

This results in negative new_core_hrs as those hours are removed, and the total_core_hrs no longer reflects the total core hours as the shutdown manager has been removed.

That makes these values somewhat awkward to use.

In this example, observe negative new_core_hrs and that total_core_hrs has gone back to 0:

1658831219.077176 2022-07-26 12:26:59 DEBUG MainProcess-119578 MainThread-139748935808832 funcx_endpoint.endpoint.interchange:362 _main_loop Publishing message b'\x01{"message_type":"ep_status_report","data":{"endpoint_id":"ee4737d1-cb0e-4048-93e7-b69fede8e4e3","ep_status_report":{"task_id":-2,"info":{"total_cores":0,"total_mem":0,"new_core_hrs":-0.33205335670047337,"total_core_hrs":0,"managers":0,"active_managers":0,"total_workers":0,"idle_workers":0,"pending_tasks":0,"outstanding_tasks":{},"worker_mode":"no_container","scheduler_mode":"hard","scaling_enabled":true,"mem_per_worker":null,"cores_per_worker":1.0,"prefetch_capacity":10,"max_blocks":1,"min_blocks":0,"max_workers_per_node":Infinity,"nodes_per_block":1}},"task_statuses":{}}}'

This is because each iteration, the total_core_hours is calculated by summing over the currently active managers:

        for manager in self._ready_manager_queue:
        ....
            core_hrs += (active_dur * total_cores) / 3600

To Reproduce
Watch an endpoint report ep_status_report before and after a manager is shut down.

Expected behavior
The total_core_hrs field should report the total core hours.

Environment
funcX endpoint main at fd9c2cf