CBT lately had 2 bugs that the backup vendor had to address.
After increasing disk space of an existing vmdk it may happen that the CBT is broken.
So take care that your backup software do not only take an active full but also reset CBT on the fly till VMware fixed that and you are running on that patch level. (Veeam does this for example. Patch was available very early after the VMware bug was up.)
But before you really go and do a Active Full for everything, let´s discuss the down side and why VMware invented this.
If you want to use VMware VADP API for Backup (or old VCB) you need to create VM Snapshots and hold it for backup.
The longer these Snapshot are present, the more data is in it and the more load it brings at Snapshot commit.
Snapshot commit basically hold on the VM and write a bunch of data to the disk file and release it afterwards. It will do this till all data was written to the right files. There are some situations that VMware is to slow at this process and can not keep on with the new written data that came in from the Applications.
Anyway in nearly all cases the Snapshot release works good.
If you have a slow storage system, Applications with Heartbearts (e.g. Exchange DAG Cluster) or Jet DB Streams this can be problematic.
So Veeam decided to add a lot of features that helps in that space like Parallel processing of VM disks, Storage Snapshot Integration, Snapsht Hunters, IO Control and others so that the VM snapshot lifetime is as short as possible.
As you Application uptime matters a lot, the use of CBT is key here.
Veeam has an additional approache. SureBackup boots an VM after a backup in a Sandbox(VirtualLab) Environment and checks if the OS is able to boot, the Network IP is available (in the Sandbox of cause) and if your selected applications can run. All this automatically.
This give you a really stable and trustfull situation while using all the good things from CBT.