This is gonna add much complexity, and needs lots of change in database structure.
Even if this is possible, it's not worth this much change and complexity.

Yes, it needs a lot of changes, but there may be an easier solution
I'm trying to find it

First thing is that, we don't need to create many UUID for each user, we can just append the Outbound name to user's UUID and then change this string to a new UUID! (user's UUID will become a master UUID)

import uuid

def str_to_uuid(input_string):
    return uuid.uuid5(uuid.UUID('00000000-0000-0000-0000-000000000000'), input_string)

print( str_to_uuid("4d0eb5f1-6bb0-44a7-8ee8-eafdc1e1c933_warp") )

or we can directly use user's UUID as namespace of UUIDv5

import uuid

print( uuid.uuid5(uuid.UUID('4d0eb5f1-6bb0-44a7-8ee8-eafdc1e1c933'), 'warp') )

for Trojan Inbounds, can just append outbound name

First thing is that, we don't need to create many UUID for each user, we can just append the Outbound name to user's UUID and then change this string to a new UUID! (user's UUID will become a master UUID)

import uuid

def str_to_uuid(input_string):
    return uuid.uuid5(uuid.UUID('00000000-0000-0000-0000-000000000000'), input_string)

print( str_to_uuid("4d0eb5f1-6bb0-44a7-8ee8-eafdc1e1c933_warp") )

or we can directly use user's UUID as namespace of UUIDv5

import uuid

print( uuid.uuid5(uuid.UUID('4d0eb5f1-6bb0-44a7-8ee8-eafdc1e1c933'), 'warp') )

for Trojan Inbounds, can just append outbound name

This part is not a big problem, the real problem is the record usage job, we're gonna have multiple records for each user instead of 1, also this gonna be applied for each node

record usage job only uses user-id and ignores rest of the email, so we can just add outbound name after the username and there will be no problem

record usage job only uses user-id and ignores rest of the email, so we can just add outbound name after the username and there will be no problem

Marzban/app/jobs/record_usages.py

Lines 104 to 110 in 349451d

	def get_users_stats(api: XRayAPI):
	try:
	params = defaultdict(int)
	for stat in filter(attrgetter('value'), api.get_users_stats(reset=True, timeout=30)):
	params[stat.name.split('.', 1)[0]] = stat.value
	params = list({"uid": uid, "value": value} for uid, value in params.items())
	return params

just needs changes in add/remove/alter/update users at api operations
https://github.com/Gozargah/Marzban/blob/master/app/xray/operations.py

i think i didn't says it right, i mean job its gonna be much slower and its not efficient

The only thing that can take longer time is recording new values to database
because there will be multiple records for every single user
But this can get optimized by combining all records of same uid before writing the new data usage to the database
and not all users gonna add 10 outbounds, mostly will only add warp outbound and done, so even if it's a problem, it's for less than 5% of Marzban users

not for now , maybe in future.