get_services() normalised table names through Services.get_tag() and then handed the result back as an identifier, so kv copy and sync skipped any table whose name differed in case or matched an alias. Yield stored names, and resolve tables case-insensitively.
