How Netflix Uses Chaos Monkey and AWS to Build Resilient Cloud Services

Netflix Overview

Netflix started as a DVD‑rental service founded by Reed Hastings after a frustrating experience with overdue fees at Blockbuster. By moving to a subscription model and eliminating late fees, Netflix grew rapidly, eventually outcompeting Blockbuster and becoming a global streaming leader with over 81 million subscribers.

In 2006 Netflix launched its streaming ("streaming") service, coinciding with the rise of broadband and YouTube. Subscription numbers jumped from 4.2 million in 2005 to 83.2 million a decade later, illustrating the cost‑advantage and scale benefits of an online delivery model.

Netflix’s success is not only due to content and business model but also to heavy technical investment. The company runs its entire IT stack on Amazon Web Services (AWS) and openly shares its engineering practices on the Netflix Tech Blog.

Chaos Monkey and the Simian Army

To ensure reliability at massive scale, Netflix created Chaos Monkey, a tool that randomly terminates production instances during business hours, forcing engineers to design systems that can survive unexpected failures. This concept expanded into the Simian Army, a suite of “monkey” tools that inject various faults:

Chaos Monkey – randomly kills instances.

Latency Monkey – adds artificial delay to REST calls.

Conformity Monkey – shuts down instances that violate best‑practice rules.

Doctor Monkey – removes unhealthy instances.

Janitor Monkey – recycles unused resources.

Security Monkey – scans for security misconfigurations.

10‑18 Monkey – checks localization and internationalization settings.

Chaos Gorilla – simulates an entire AWS Availability Zone failure.

Netflix and AWS

Netflix is one of AWS’s most important customers, accounting for roughly one‑third of North American internet traffic that passes through AWS. The partnership is symbiotic: Netflix relies on AWS for compute, storage, big‑data processing, and AI services, while AWS showcases Netflix as a flagship case study at events like re:Invent.

After a 2008 database corruption incident that caused a three‑day outage, Netflix decided to outsource its infrastructure to AWS, embracing cloud elasticity, rapid scaling, and automated deployment tools. This move spurred the creation of many open‑source Netflix OSS projects such as Eureka, Ribbon, Hystrix, Genie, Spinnaker, EVCache, and the Simian Army itself.

Key Takeaways

Operating at massive scale on a shared‑resource cloud requires designing for failure: “the best way to avoid failure is to fail constantly.” Real‑world testing, continuous chaos experiments, and a culture that embraces antifragility enable Netflix to maintain high availability and rapid innovation.

By deliberately exposing systems to chaos, Netflix turns potential weaknesses into strengths, embodying the principle that “what doesn’t kill you makes you stronger.”