Google Cloud Platform for Developers

Autoscaling load balanced resources

All types of Google load balancers support managed instance groups (MIGs). As we've seen in Chapter 7, Google Compute Engine, one feature of managed instance groups is that they support autoscaling. One or more autoscaling policies define a metric (CPU utilization or Stackdriver metric) and target utilization. When the target utilization is reached, the MIG will add additional compute instances to the group based on the instance template. HTTP(S) load balancers extend this functionality by adding support for scaling based on load balancing serving capacity, as defined on the backend resource.