Introduced this to be able to monitor how big the numbers actually are and if we need to worry about performance

Decided to do a traditional for loop instead of spawning threads, to avoid issues similar to our last OOM incident.
However, if the sequential query becomes too slow, I can try splitting this into chunks and spawn threads for each chunk of let's say 5-10 requests

Unfortunately, we cannot do anything like BatchGet here because we're querying an index.

I hope this will limit the number of queried user IDs over time. The more users registered on Identity we have, the less users will stay in the reserved usernames table