The best feature I like is the ability to do chain slave replication, to alleviate some of the load on the master. Its also very easy to integrate, regardless of the language you choose.
Another very useful feature of Redis is the ability to do set theory computation easily.
There is very little to dislike about Redis. The only thing I can complain about is clustering. Its not very easy to setup a cluster but once you get the hang of it, then it becomes a bit easier.
If your considering a caching solution or some number crunching. Redis is perfect.
We used Redis to do a all of our set theory computation on data we retrieve from social media. With the built in sets, it was a perfect off the self solution.