Bek
It's nginx cache, and honestly not the most ideal, but much more straight-forward than a suitable alternative. I'm not ready to make the configuration public, as we're still testing and it could prove to need some tweaks before being without fault.
One of the (many) cool side-benefits of this caching system is that the site will continue to dish out cached content in the event of a server misconfiguration that would otherwise display an error 500 message. I've tested this and it works like a charm! In production we will have failover servers, and public-facing proxy servers that handle SSL termination and caching. In the case of a webserver fault, people would see cached content from the proxy server before being redirected to the backup. The end user should experience no downtime, save the inability to post content for that minute. Really excited to test this out in the coming weeks!
