Configure EC2 Auto Recovery for SAP Instances

Quick Answer

EC2 Auto Recovery is a feature with specific mechanics, and this question is testing whether you know what actually has to be in place for it to work rather than what would generally be desirable for availability. The trigger mechanism is a CloudWatch alarm watching the StatusCheckFailed metric - when that check indicates the underlying hardware is impaired, the alarm is what initiates the recovery process automatically. Beyond that trigger, the instance itself has to be explicitly configured for EC2 Auto Recovery as a prerequisite; it isn't a behavior that applies to every EC2 instance by default. The third required piece is understanding a specific constraint of the feature rather than an additional component to add: EC2 Auto Recovery operates within a single Availability Zone, recovering the instance onto new underlying hardware in the same AZ it was already running in, rather than failing it over to a different AZ. That's an important distinction from HA architectures like Multi-AZ database deployments or cross-AZ clustering, which are designed specifically to survive an entire AZ going down - Auto Recovery instead addresses the case where the specific hardware host beneath an instance fails, while the AZ itself remains available. An Elastic IP isn't a required component either; if one is already associated with the instance, it gets remapped automatically as a side effect, but it isn't part of what needs to be configured to enable recovery. The pattern to remember is that Auto Recovery is a same-AZ, hardware-failure remedy, not a cross-AZ high-availability mechanism.