[CBM] Clarify CPU usage increases with number of KV nodes

What is the problem?
The --threads parameter is a little difficult to understand as it is not really the number of threads, it's the number of workers we use to do a backup. Each worker will transfer a vBucket per KV node, which can lead to high CPU usage when customers use a small number of threads but have a very large cluster

What is the solution?
We should modify the documentation (maybe that of threads) to clarify that the cluster size can affect resource usage, as well as the number of threads.