The least connections load balancing method functions best when all pool members share similar characteristics.

The least connections algorithm treats active connection count as a direct measure of server load, which is only an accurate proxy when every pool member has comparable CPU, memory, and application throughput capacity. When members are homogeneous, the server with the fewest connections is genuinely the least loaded - but if capacities differ, a lower-spec server with few connections may be more saturated than a higher-spec server handling many, making the metric misleading.

False is incorrect because the least connections method's fairness assumption depends on all pool members being able to handle connections equivalently, which is precisely why similar characteristics across members are a documented best-practice requirement for this method.