Least Connections

A dynamic load balancing algorithm that routes each new request to the backend server currently handling the fewest active connections. Unlike round-robin, which ignores connection duration, least-connections accounts for variable request processing times and is particularly effective when requests have widely differing execution durations (e.g., long-running WebSocket connections mixed with short API calls). A weighted variant — Least Connections with weights — adjusts the effective connection count by the server's assigned weight, combining capacity awareness with dynamic load measurement.