The lights went out in the data center, but your system should keep running. Least Privilege Chaos Testing is how you find out if it will.
This discipline combines the principle of least privilege with chaos engineering. It isolates the real consequences of over-privileged access before an attacker or accident does. Every permission, token, and role is tested under stress conditions. You remove excess privileges, then break things on purpose to measure the result.
The principle of least privilege limits users and services to the minimum access needed. In practice, teams often grant more permissions than required, chasing speed or avoiding friction. Over time, this creates a broad attack surface. Silent security debt builds, waiting for exploitation. Least Privilege Chaos Testing forces that debt into the open.
A typical test revokes a critical API permission mid-operation. Or replaces a stored key with a restricted one while services are live. You observe what fails, what degrades, and what alerts are triggered. These are not theoretical drills. They simulate genuine conditions: expired credentials, role changes, compromised keys, partial system failures.