Logstash horizontal autoscaling

I want to create autoscaling for our logstash based on prometheus queries we have from the api of logstash.

I am using Persistent Queue has a way for backing up our messages in case of disaster and I thought that based on the ammount of usage of our queue I could see if we should auto scalie (used_bytes/max_bytes).

I wanted to know couple of things before:

  1. how do I decide the prefered size of logstash persistent queue (I think mine is pretty large cause normally it has utilization of 0%) ?
  2. what are other metrics I should probably use to auto scale (I thought regarding maybe amout of events in divided by the amount of events out and to see if it is bigger than 1.1)

also if someone implemented this kind of thing it will be helpful