It is an algorithm similar to Round Robin but the difference is that it takes into consideration the capacity and assign more requests to the server that can handle more requests.
How it works
Weighted round robin redirects requests based on the load. In the example below, heavy loads go to the server 1 that has higher specification, because it is more powerful and is capable of processing it. The low lecel can go to Server 2, which has lower spec.
Transclude of Weighted-Round-Robin.excalidraw
For the load balancer to understand where to point to, you have to assign the weights to each node. In the case above, we said that the first 2 will go to Server 1 and the next one will go to Server 2, and then it cycles back to Server 1.