AWS – Amazon EC2 Auto Scaling Warm Pools now supports hibernating instances and returning instances to Warm Pools on scale-in

Amazon EC2 Auto Scaling Warm Pools now support two new features: you can now hibernate your Warm Pool instances and you can configure your Auto Scaling group to return running instances to a Warm Pool on scale-in. With Warm Pools, you can scale out faster by bringing instances into service from a pre-initialized pool of EC2 instances. Warm Pools are a good fit for applications that have time consuming initialization steps — like loading gigabytes of data, provisioning services, or running custom scripts — that can take several minutes or longer before those EC2 instances are ready to serve traffic.

Hibernate Support for Warm Pools helps applications with significant memory state scale-out faster. You can now choose to put your Warm Pool instances in the Hibernated state, in addition to the previously supported states of Stopped and Running. Choosing hibernation allows you to pre-initialize the entire EC2 instance state, not just the disk state, for instances that need to rebuild significant memory state before being placed in service.
Scale-in to Warm Pool saves you from having to rebuild a new Warm Pool instance during scale-in. This enhancement allows running instances to be placed back into the Warm Pool during a scale-in event. Previously, EC2 Auto Scaling would terminate one of the running “in-service” instances during scale-in and then launch and pre-initialize a new instance to replenish the Warm Pool. Now, instances can be placed back into the Warm Pool in the desired state – stopped, hibernated, or running, but idle.

Read More for the details.