Chaos Engineering

What it is

Chaos engineering deliberately injects controlled failures (killing an instance, adding network latency, exhausting a resource) into a system — usually in production or a production-like environment — to verify resilience assumptions actually hold.

Key points

• The core motivation: most systems have untested resilience assumptions — a documented failover that was never actually triggered end-to-end often doesn't work when it's needed for real.
• Experiments start small and contained (a single instance, a small traffic percentage) with a clear rollback plan, not by blindly breaking production.
• A blast radius — the scope of what's allowed to be affected — is defined and enforced before every experiment.
• The output of chaos engineering is typically a list of found gaps (a missing retry, an untested failover path) — it's a discovery tool, not a one-time certification that a system is now resilient.