The logic of the GP still applies though. Sites have outages every day so it is inevitable that some large sites will fail around the same time. Also, we know that Cloudflare and Twitter outages were attributed to configuration changes, probably others have benign explanations as well.