archive coverage:

• the number of tracked "forges": this include singleton forges (e.g., GitHub) and each instance of self-hostable forges (e.g., GitLab). It's of course a very crude metric, but why not
• the total number of origins
• the total number of "dead" origins, i.e., that haven't seen an update in more than a given threshold (this metric can be interpreted in various ways, depending on our backlog, but it's useful either way)