Git.kernel.org receives approximately six million daily requests, most of which are automated crawlers that render git commit history as HTML. This process consumes significant CPU resources, with 14 of 90 total cores dedicated to rendering commits, representing about 20 % of the system’s capacity. Anubis proof‑of‑work challenge blocks roughly two‑thirds of the requests; the remaining requests succeed after solving a difficulty‑based math puzzle. Legitimate user traffic is estimated at two % of total requests, while the majority are scrapers seeking the full commit history for training large language models. The project has disabled certain features and limited anonymous access to reduce expensive URL exposure and computational load.