Would be interesting to share the different in ops cost (year support paid aside), for example, do you need a beefier server now? more servers? less? size on disk? IOPS of servers? also, any different in your metrics from an end user perspective? e.g: time to publish an article.