Same goes for MySQL. I like to try new and different things, so I'm always trying to find reasons to use things like redis or other pub/sub or caching options. But I usually tend to end up sticking with tried and true relational DBs. Obviously, there is a scale at which those other options will be necessary, but I haven't hit that yet.
Many times even when I do use something like Redis, I run into limitations.
For example, I have been using it for caching calculated data that users need, but now we need to be able to query the cached data by date which puts us back to needing to store the cached data in the relational database.