Check your total of the target_size_bytes, it could be bigger than the "Raw Capacity" available. If not then the BIAS can play a role as well, which acts as a multiplier for increasing the number of PGs. In short, the autoscaler assumes in the future you may be using more space then the raw capacity.
Have a look at the docs for the autscaler:
https://docs.ceph.com/en/latest/rad...nt-groups/#viewing-pg-scaling-recommendations
I recommend to use the "Target Ratio" instead of the "Target Size" which will use a floating ratio relative to the pools on the same crush device class.