Cloning a Git repository is expensive, as the entire repository history must be transferred. Clients that have cloned a repository before should rather fetch the latest changes. This is (sometimes dramatically!) faster for the client and less expensive for the server.
High numbers of clone operations usually indicate an unfavorably configured CI process.
The amount of data transferred for clones and fetches (excluding Git LFS traffic) as well as pure Git LFS traffic (indirectly triggered through clones and fetches).
To save resources, you should try to minimize the number of clones, especially for large repositories.
This table lists the repositories that accounted for the majority of the traffic yesterday.