Yeah, we're using Cassandra for logging. Not quite as simple to scale up, but it's write-only in the request cycle and hasn't been anywhere near a bottleneck yet.