Hello,
Being new to Azure and many things Cloud, I am seeking an answer to the behavior of Azure as it relates to autoscaling virtual machines. An example and then the question.
Example:
I have 4 VMs in an availability set, with various application endpoints created as load-balanced sets. I then enabled autoscale by metering CPU use for my availability set. When overall CPU load rises or lowers based on the thresholds, additional VMs are started or stopped, although this CPU load may not be due to the active sessions to the load balanced endpoints.
Question:
When stopping a VM due to a lower threshold being met, does Azure simply stop the VM regardless of active endpoints sessions or is there some type of mechanism that delays the stop until no more sessions are active to that VM?
Thank You.