a few hundred thousand repositories of the 85 million reported Github repositories is really not much, see https://github.com/ten at the bottom for the number.
of those 85 million how many are actually active and not just placeholder empty repos or automated mirrors for wordpress/jquery plugins? Not on anyones side here but if the top 250k repos switched, that would be really bad news. I think it's more important to compare size and popularity of the repos than just what the % is.
i assumed that only small repositories that nobody except the author is using can be moved in such a small timespan of a few hours. Heavily used open source projects have issues indexed in Google with permalinks, their users probably hardcoding git urls, connections to package managers, heavily customized Travis CI integrations and the like.
Of those that switched, how many were backup mirrors vs actually switching to the Gitlab platform?
We can only guesstimate with the data we have.
Yeah. That has to go both ways too. If it’s only tiny repos moving over and not top repos.