Wish they'd do a full write-up. I'd be interested in how they run their node processes.<p><pre><code> - Using containers, VMs, bare metal?
- Using a process manager like Forever[1], PM2[2] or just native systemd?
</code></pre>
[1] <a href="https://github.com/foreverjs/forever" rel="nofollow">https://github.com/foreverjs/forever</a><p>[2] <a href="https://github.com/Unitech/pm2" rel="nofollow">https://github.com/Unitech/pm2</a>
I'm still confused.<p>Did they add a layer between their backend and mobile to pre-render the page server side, or did they get rid of their whole scala/jvm stack to replace it with Node?
They seem to have done this only for Webpack (splitting JS libs + prefetching).<p>Why use Node at all? It's all static content...<p>Crazy that a top engineering post is about some company discovering "compile-time".<p>Next post will be about how they discovered this thing they call Testing?
I remember seeing Twitter switching from ruby/rails to the JVM for a lot of their backend a few years ago. I wonder why they're using node now (unless the mobile traffic was never moved to the JVM)?<p><a href="https://www.infoq.com/articles/twitter-java-use" rel="nofollow">https://www.infoq.com/articles/twitter-java-use</a>