One of the key features of btrfs is the snapshot and roll back capability. However in a Cloud framework, Azure, EC2, GCE, OpenStack, root volumes are generally small and snapshots are generally taken externally, i.e. tooling provided by the framework. Given a small root volume snapshots internal to the filesystem, such as the btrfs snapshots, take up space that on a percentage basis is more valuable than on a physical server where the root volume may be 100 or more GB. Further, while in Azure and many other frameworks the boot menu can now be accessed via serial console that implies that GRUB boot timeout has to be set to a sufficient amount of time to get there, probably on the order of 45 seconds or more. This implies that every boot would be delayed by that amount of time whether one wants to boot into a previous snapshot or not. The btrfs feature of internal snapshot and roll back is more or less superseded by the features provided by the Cloud framework.