Techment — Site Header

What is Chaos Engineering

Definition

Chaos engineering is the practice of deliberately injecting failures into a system in a controlled way to test its resilience, revealing weaknesses in how it handles outages, latency, and faults before real incidents expose them in production.

« Back to Glossary Index

Reveals hidden resilience weaknesses before real outages do
Builds confidence that systems survive realistic failure conditions
Validates recovery, failover, and redundancy actually work
Drives a culture of designing for failure rather than assuming stability

Real World Example

A streaming company runs a chaos experiment that randomly kills service instances in production, discovering that one dependency lacked proper failover, and fixes it before a real node failure could cause a customer-facing outage.

FAQs

What is the goal of chaos engineering?

To proactively expose and fix resilience weaknesses by deliberately injecting controlled failures before real incidents occur.

Is chaos engineering done in production?

Often yes, carefully and with safeguards, since only production reveals true system behaviour under failure, though it can start in staging.

What kinds of failures are injected?

Instance termination, network latency, resource exhaustion, and dependency outages are common chaos experiments.