Wikis with more than 1 million rows in revision may have issues with editcounts by namespace, since the extension is doing a count(*) from revision providing the user ID and a join to page to get the namespace.

Wondering if it may be saner to have a new table for this extension, with the aggregate of counts by namespace, and update this table on each edit, as it's done with the editcount from the user table, to avoid to scan a large quantity of rows...