We have an issue in counting metrics for Tor bridges where we are not accounting for nodes that use the same fingerprint but different nicknames (#40047 (closed)). From what it has been estimated we are undercounting bridge users by about 14,000 or 14%, since only 12.5% of Snowflake users are being counted. (onionoo#40022 (comment 2793712))
This is an issue with our assumption that the fingerprint is the ultimate identifier for a tor node. This MR introduce the nickname when counting clients.
While this won't be able to show different graph for each group of nodes with the same fingerprint but different nickname it would still correct our aggregated statistics that we publish on metrics.torproject.org