To fully answer that I guess I'd need a clearer idea of what environment you're hoping to construct. Cloud Load Balancers does allow for failover from a primary to a secondary server, along with weighted load balancing. You can see some details of what options are available when adding a node in our API documentation:

http://docs.rackspace.com/loadbalancers/api/v1.0/clb-devguide/content/Add_Nodes-d1e2379.html

Most recovery options for Cloud Servers you would need to implement yourself, generally speaking. You can schedule backups with Rackspace Cloud Backup, for example, and then use its API for full control over both backups and restores:

http://docs.rackspace.com/rcbu/api/v1.0/rcbu-devguide/content/overview.html

A tool like chef or puppet would let you combine that with a quick means of deploying a replacement or additional server for an environment. You could also set up a private Cloud Network behind a load balancer and use Linux-HA to handle your own failover directly. And of course you wouldn't need to use our load balancer product - you could set up your own reverse proxy like varnish to handle load balancing from a Cloud Server.

If you opt for Cloud Sites, you'd have to handle backups on your own, but we would handle the HA for you. It's a shared web environment, so you have less control over the environment configuration but have to worry less about dealing with availability concerns outside your own web application.