There are many tools available for building monitoring systems. Nagios is just one of the common general-purpose monitoring tools. As long as some application is checking on the health and availability of your system and letting you know when an issue is present, or about to present itself, you will be in good shape. There are also some good examples of how your main application and other parts of your application interact with Cassandra and can be instrumented to give you a feeling of total information awareness and potentially the ability to get a good night’s sleep when it is all in production.
In this example, Nagios can act as an early-warning tool. It can give you a heads-up to look at the machine in question and dig deeper into a potential problem before it turns into something more serious such as a full-fledged outage or a completely downed node. Ensuring an intelligently set-up monitoring infrastructure is essential to having a well-designed and architected system.