Two identical servers in one room, two communication links on the same physical route, or two backup power sources dependent on one node can create the appearance of redundancy without real independence.
Common-mode failure
Systems analysis should look not only for weak components but also for shared dependencies: power, software components, physical routes, personnel, external suppliers or a single authorization process.
Distribution reduces the risk of complete shutdown but adds coordination complexity. It therefore needs clear rules for local autonomy and resynchronization after connectivity is restored.