The Compute Engine feature that allows virtual machines (VMs) to be automatically added to or removed from an application based on load metrics is called autoscaling. Specifically, autoscaling is managed through Managed Instance Groups (MIGs) in Google Cloud Compute Engine
. Autoscaling works by monitoring defined metrics such as average CPU utilization, HTTP load balancing serving capacity, or custom Cloud Monitoring metrics. Based on these metrics, the autoscaler automatically increases (scales out) or decreases (scales in) the number of VM instances in the group to match the current demand, ensuring efficient resource use and cost savings
. Key points:
- Autoscaling is configured via a policy that specifies which metrics to monitor and the desired target utilization.
- The autoscaler continuously compares actual usage to the target and adjusts the number of VMs accordingly
- You can also set up schedule-based autoscaling to prepare for anticipated changes in load
In summary, Compute Engine autoscaling for Managed Instance Groups is the feature that dynamically adjusts the number of VMs based on load metrics
