Can you share the metrics databricks uses during job cluster auto scaling?
Is Databricks  looking at queued tasks, slot utilization etc or just looking at CPU utilizations?

The autoscaling docuemnt https://docs.databricks.com/aws/en/compute/configure?utm_source=chatgpt.com#how-autoscaling-behaves
doest provide details around this.

Also looks like the enhanced autoscaling is not available on job cluster