Octavia GET request for members in pool takes a lot of time on huge Loadbalancer

Create LB with big topology, f.e.:
39 listeners with 1 pool and 13 members in this pool.

so totally it will be about 600 objects in the objects graph:
39 Listeners, 39 HMs, 39 pools, 13*39 members, LB, VIP, etc...

Such setup allows to display performance issue with GET requests.
Specifically when I try to get information about members in pools, i.e. call:
GET /v2/lbaas/pools/<pool_id>/members/

This call takes around 2.5 seconds.

GET /v2/lbaas/pools/78b285be-48fa-4470-a838-8445ddf83774/members => generated 6893 bytes in 2390 msecs (HTTP/1.0 200) 4 headers in 158 bytes (1 switches on core 0)

After profiling code for handling GET request, I found, that around 1 sec on call:
model.to_data_model()

https://github.com/openstack/octavia/blob/master/octavia/db/repositories.py#L174

Looks, like it happens due to recursive iteration through all nodes in graph (https://github.com/openstack/octavia/blob/master/octavia/db/base_models.py#L59)

Probably this could be optimised by only one step of recursive call, i.e. do not go through all nodes in the graph (for member it's only 1 pool, instead of all 600+ nodes in graph).

During local testing on the same topology with only one recursive call, it took around:
1.2 sec:

GET /v2/lbaas/pools/78b285be-48fa-4470-a838-8445ddf83774/members => generated 6893 bytes in 1432 msecs (HTTP/1.0 200) 4 headers in 158 bytes (1 switches on core 0)