From what I gathered from Michal Slaski’s presentation, this is not a tool for production, but only for peering into your project in your dev environment. FYI. At least as of ~March 2017.
tracing in erlang is cheapish but non-free, and bringing it online in production is one of those things you have to be super careful about. Scalpels like eper/redbug (https://github.com/massemanet/eper) and recon_trace (http://ferd.github.io/recon/recon_trace.html) offer a little more control and discrimination.
Netflix seem to use the same visualizations for their Chaos Monkey project.
I could find back the article here: https://medium.com/netflix-techblog/flux-a-new-approach-to-system-intuition-cf428b7316ec
The fail-over video is really impressive.