On AWS I've always imaged machines that contained SQL Server installations and running databases. When I spin up new instances from the images, SQL Server and all database always returned without issue (e.g. consistency checking always came back successful).
Is this to be expected? i.e. SQL Server being ACID compliant, allows a snapshot of the disk to be taken due to the data in the DB file and log being in a consistent and durable state, so allowing the database to recover itself when it comes back to life when spun up on a new instance from the image.
The following seem to support this idea, but discussion at work this week have resulted in me questioning this?
https://docs.aws.amazon.com/prescriptive-guidance/latest/backup-recovery/ec2-backup.html
Does anyone have any definitive information, and if this is ok to do, is it AWS only, or do other IAAS platforms (Azure / GCP) support this as a DR strategy?