Simple definition
Fault injection deliberately introduces controlled failures to test system resilience, recovery, and operational safeguards.
Purpose and use
Fault injection creates controlled failures to observe system behavior and test recovery mechanisms.
How does it work?
An experiment defines a hypothesis, limited targets, stop conditions and measurements, for example when a dependency becomes unavailable.
Example and verification
Start with a suitable environment and scope, then verify that alerts and procedures actually detect and handle the simulated incident.
Points to watch
Incorrect targeting can interrupt service. Absence of visible failure does not prove coverage of every resilience scenario.
In short
Value comes from an authorized, bounded and measurable experiment followed by action on observed weaknesses.