Hi Dotex,
What kind of limit do you mean?
I would spread backup operations in time. Worst case scenario read operations will be queued but that does not help with the read of the data nonetheless.
Also, I would make certain that it is the same snapshot file that is being replicated or that the snapshot files are rotated frequently enough to avoid piling up and cause latency on the parent VM. From what I know a backup solution normally does a call to take a snapshot first and then copies it across to the desired destination. The more snapshots a VM has (the longer or deeper the snapshot tree is) the higher is the latency to read/write from/to the VM’s disk. Therefore it is beneficial to keep your snapshot list for a VM short and fresh i.e. with no old snapshots.
Please let me know if that is helpful.