Skip to content
BENIBIX Learn

Chaos Engineering

Infrastructure📟 ReliabilityResilient Systems

Also known as Chaos Testing, Fault Injection, Chaos Experiment, Chaos Monkey

Breaking parts of a system on purpose, to find weak spots first.

Think of it like

Think of it like a fire drill. You ring the bell on purpose, so everyone knows what to do in a real fire.

Example

Lucy's team stops one of three web servers in a quiet hour. Orders keep coming in. So the load balancer really does go around a dead server.

Why it matters

Many systems are built to survive faults that nobody has ever tried. These tests check that the plan works.

Definition

Chaos engineering is breaking parts of a system on purpose to learn whether it copes. The team first writes down what normal looks like, such as orders per minute. They guess that normal will hold, then cause a real fault: stop a server, cut a network link or slow a database. If normal does not hold, they found a weak spot before it found them. Tests start small, so one going wrong hurts few users.
This loop is the method: write down normal, break one part, and see if normal holds.

Common mistake

Starting big, in the busiest hour. Begin with one small fault, a way to stop at once, and few users in reach.

Key takeaways

  1. 1
    Normal first. Write down what normal looks like.
  2. 2
    Break on purpose. Stop, cut or slow a real part.
  3. 3
    Start small. Keep the harm to a few users.