Skip to content
Latchkey

What Is Chaos Engineering?

Chaos engineering is the disciplined practice of injecting controlled failures, such as killing instances, adding network latency, or exhausting resources, to verify that a system tolerates them gracefully. Experiments start from a hypothesis about steady-state behavior and observe whether the system holds it under stress. The goal is to find brittleness before an unplanned outage does.

Why it matters

Resilience features like retries and failovers are often untested until a real incident exercises them, and then they sometimes do not work. Chaos experiments validate those defenses on purpose, in a controlled blast radius, so weaknesses are fixed proactively rather than discovered at 3 a.m.

Related concepts

  • Experiments are bounded by a small blast radius
  • Validates circuit breakers, retries, and failover
  • Popularized by tools like Chaos Monkey

Related guides

Tired of flaky CI? Latchkey auto-heals failed jobs and retries them for you. Start free →