Surprisingly, Netflix Didn’t Go Down During the “Conformity Gate” Traffic Surge. Here's Why

Why Netflix Didnt Go Down During The Conformity Gate Traffic Surge
Follow Us:
1k
1k

The fast-spreading fan theory dubbed “Conformity Gate” had thousands of Netflix users logging on around 8 p.m. Eastern time on Jan. 7. They all expected a surprise “Stranger Things” finale, but nothing new ended up coming out. Still, the rush to the platform was ever-present.

And yet, Netflix didn’t fold under the traffic pressure. It’s impossible to know exactly how much Netflix actually saw that night, but most users appeared to experience little more than brief slow-loading.

While Downdetector received about 1,400 reports, they were scattered between 7:55 p.m. and 8:10 p.m. Eastern time, and there were no spikes that were actually sustained and nothing official from Netflix’s status page.

Netflix outages reported in the last 24 hours - Downdetector
Source: Downdetector

The expectation that Netflix would have crashed isn’t totally irrational. “Stranger Things” pulled in 59.6 million views within five days of the premiere of its final season. When that much traffic shows up at the same time, crashes tend to feel…inevitable.

We reached out to Netflix for comment, but did not receive a response by the time of publication. Still, the company has been fairly open over the years about how its infrastructure is built.

So, Why Didn’t Anything Happen?

Netflix’s preplanning and infrastructure purposely helped avoid something more catastrophic.

To start, Netflix operates its own proprietary global CDN, Open Connect, which uses edge caching, predictive capacity planning, and cloud autoscaling.

Diagram of Netfli'x PlayAPI
Netflix gives guaranteed priority to playback requests, while background pre-fetching is allowed only if the system has spare capacity. Source: Netflix Tech Blog

Netflix’s architecture also splits control-plane services in the cloud (like account and metadata requests) from the data plane that streams video via edge caches, so it isolates and scales different workloads separately.

Its engineers also use Chaos Monkey, which intentionally injects issues into production so systems learn how to react to specific issues or behaviors.

Also, you have to consider that not every traffic request hits Netflix the same way.

During busy spikes, Netflix uses a method called “prioritized load shedding.” If the system sees an unusual amount of activity, it protects the information actually needed to start playback, and gets rid of the less important background work first, like browsing-related preloading and previews.

While Conformity Gate likely caused a surge in activity, most of that extra load would have landed on the background work mentioned.

And that’s why thousands of users barely noticed a hitch. For the most part, at least.