Instance resiliency - Amazon Elastic Compute Cloud

Instance resiliency


The following information applies to configuring recovery-related capabilities on healthy instances. If you are currently encountering difficulties accessing your instance, see Troubleshoot EC2 instances.

In the event that AWS determines an instance is unavailable due to an underlying hardware issue, there are two mechanisms that you can configure for instance resiliency which can restore availability—simplified automatic recovery and Amazon CloudWatch action based recovery. This process is called instance recovery.

At least one mechanism must be configured or enabled in advance with supported resources for the instance recovery process to occur. By default, simplified automatic recovery is enabled for supported instances when they are launched.

Overview of instance recovery

The following are examples of underlying hardware issues that might require instance recovery:

  • Loss of network connectivity

  • Loss of system power

  • Software issues on the physical host

  • Hardware issues on the physical host that impact network reachability

A recovered instance is identical to the original instance, including its:

  • Instance ID

  • Public, private, and Elastic IP addresses

  • Instance metadata

  • Placement group

  • Attached EBS volumes

  • Availability Zone

A successful instance recovery will appear to the instance as an unplanned reboot. In other words, content stored in volatile memory will be lost, instance store data will be erased, and the operating system’s uptime will start over from zero.

To help protect against data loss, we recommend that you regularly create backups of valuable data. For more information about backup and recovery best practices for Amazon EC2 instances, see Best practices for Amazon EC2.

Instance recovery alternatives

The following alternatives to instance recovery can be considered when they meet the use case of your instances.

Auto Scaling groups

You can use Auto Scaling groups to allow you to group a collection of instances together for the purposes of scaling and availability. In the event an instance within an Auto Scaling group becomes unavailable, the instance will automatically be replaced (not recovered) by the Auto Scaling group. For more information, see What is Amazon EC2 Auto Scaling? in the Amazon EC2 Auto Scaling User Guide.

Amazon EBS Multi-Attach

You can configure Amazon EBS Multi-Attach for your instances to allow multiple instances to be connected to the same EBS volume. When combined with appropriate software, this allows for high availability clustering to be enabled. For an example configuration with Linux instances, see Clustered storage simplified: GFS2 on Amazon EBS Multi-Attach enabled volumes on the AWS Storage Blog.