Our Flightcontrol engineering team developed and released autoscaling improvements for new services on Flightcontrol. Existing services will continue with their current autoscaling settings.
For new services, if you set different minimum and maximum number of instances for your web servers or workers, Flightcontrol will use the following default rules for autoscaling:
70% CPU utilization
70% Memory utilization
For web services only - 500 requests/minute
The cooldown for creation or deletion of a service is 300 seconds, so your instance pool can respond to the incoming traffic before deploying or destroying instances.
All of these default settings may be changed in the flightcontrol.json configuration file.