So I have been tasked with trying to figure out how to do failovers. The situation is 1 cloud based Storage unit and several cloud based CentOS boxes. I need to setup a heartbeat between the boxes for different services and fail over in the event one of the services stops responding. Another stipulation is to be as budget friendly as possible, and if possible free.

My first thought was SNMP packet monitoring. I have just never done it. Any ideas of how I can implement this into a production cloud environment with it being as painless as possible?

Are your servers already doing clustering? If not then I would look into that. That way the fail over will be automatic and you can rest easier. here is a good starting point. http://www.howtoforge.com/high_avail...artbeat_centos it even goes into how to setup a heartbeat between the cluster nodes.

Thanks