In an auto scaling cluster environment, when an instance stops functioning or shuts down, a new instance is provisioned using the base image configured in the auto scaling policy. Each Barracuda WAF instance synchronizes the configuration with the other instances in the cluster. A new instance that gets provisioned automatically joins the existing cluster.
If you want to upgrade the cluster to the latest image (latest firmware release), it is recommended to create a new VMSS and provision the instances with the required image. Bootstrapping can be used to launch the instances with a previously used configuration. After the instances are provisioned, the traffic can be switched from the old VMSS instance to the new VMSS instance on the load balancer. For information about bootstrapping, see Bootstrapping Methods.