Chaos Engineering
Noun · Development · Origin: 2011
Definitions
The discipline of intentionally injecting failures into production systems to verify they can withstand turbulent conditions. Netflix's Chaos Monkey randomly kills production instances. The philosophy: if it's going to break, better to break it on purpose when you're ready.
In plain English: Deliberately breaking things in your live system to make sure it can handle failures — like a fire drill for your servers.
Beyond Netflix's Chaos Monkey, tools like Gremlin and LitmusChaos let teams inject specific failures: network latency, disk full, CPU exhaustion, clock skew. The goal isn't to break things randomly — it's to form hypotheses about system behavior and verify them experimentally.
Example: 'We ran a chaos experiment simulating a database failover during peak traffic. Turns out our connection pool doesn't handle failover — it just throws errors for 90 seconds.'
Source: tooling / methodology
Origin Story
Netflix's radical idea: break things on purpose to build resilience
Netflix engineers coined **chaos engineering** around 2011 when they built **Chaos Monkey**, a tool that randomly terminates production instances during business hours. The premise was counterintuitive: intentionally causing failures reveals weaknesses before they cause outages.
The concept evolved from the **Simian Army** -- a suite of tools including Chaos Gorilla (kills entire availability zones) and Latency Monkey (injects delays). Casey Rosenthal and Nora Jones formalized the discipline in the book *Chaos Engineering* (2020).
The key insight is that complex systems fail in complex ways. Testing individual components isn't enough; you need to test the system's response to real-world turbulence. Companies like Google, Amazon, and Microsoft adopted the practice, and tools like Gremlin and LitmusChaos made it accessible beyond Netflix.
Coined by: Netflix engineering team (Casey Rosenthal, Nora Jones formalized it)
Context: Netflix, ~2011
Fun fact: When Netflix first proposed Chaos Monkey, many engineers were horrified at the idea of randomly killing production servers. The tool was initially controversial internally, but after it caught several critical weaknesses, it became a point of pride.