Improve billable users query

What does this MR do and why?

• remove distinct from queries
• add limit logic to not do queries unless needed
• add index as per analysis
• add logging for notification as per https://gitlab.com/gitlab-org/gitlab/-/issues/384527#note_1202287178

queries with complex namespace id that will run against in production

users being queried	original buffer size	remove duplicate distinct(buffer size)	with limit added(buffer size)	with specialized member index	without large namespace union(use_traversal_ids_for_ancestor_scopes feature flag enabled in production)
group_member_user_ids	585	582	582	481	n/a
project_member_user_ids	645	642	642	580	n/a
shared_group_user_ids	210	207	207	138	207
shared_project_user_ids	2284	2281	2281	2183	2257

queries with gitlab-org namespace

will never run against in production since it will only run on free/private namespaces, but is a good for benchmarking

users being queried	original buffer size	remove duplicate distinct(buffer size)	with limit added(buffer size)	with specialized member index	without large namespace union(use_traversal_ids_for_ancestor_scopes feature flag enabled in production)
group_member_user_ids	32119 48ms	32116	23036	21879 31ms	n/a
project_member_user_ids	21960 44ms	21957	18145	17989 31ms	n/a
shared_group_user_ids	34623 59ms	34620	25340	23895	23905 41ms
shared_project_user_ids	48907 67ms	48904	39596	38121	38157 53ms

MR acceptance checklist

This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.

• I have evaluated the MR acceptance checklist for this MR.

Related to #383264