I've not used this specific tool, but in general I think the things it tries to do are useful:

• Connection pooling - in my experience, this can really help performance, especially if the database is on another machine.
• Replication - can certainly help reliability (might be overkill for small projects, but if you can afford the double hardware...).
• Load balancing - it should take a while before you reach the load where this is needed, and it only works if most of your queries are selects, but eventually it's useful.
• Caching - also depends on your usage pattern (and memory availability), but can help performance in some cases.

This is based on the same concepts with other tools.