Flow metrics based auto-scaling

In contrast to CPU based auto-scaling where scaling decisions are made based on infrastructure utilization, flow metrics based auto-scaling uses back pressure prediction, data queued to determine when to scale up or down.

When you turn on flow metrics based auto-scaling, CDF uses a stabilization window of five minutes before initiating any scale up or down operation.

Use cases

Flow metrics based auto-scaling can be used to detect the need to scale based on flow performance metrics. In cases where queues fill up without CPU utilization going over the threshold, flow metrics based auto-scaling will trigger a scaling operation. Flow metrics based auto-scaling can also detect the need to scale quicker than CPU based scaling by looking at predicted back pressure as an early scaling indicator.