This is entirely true. We are big users/consumers/fans of Cassandra at Spotify. We have approximately one crap-ton of Cassandra clusters with several metric crap-tons of data. But we have an entire team that provides support and tooling around Cassandra. We contribute upstream, and have employed core Cassandra contributors in the past.

In return we get a datastore that scales pretty much infinitely with our data sets, has performance characteristics that we are now well aware of and are able to reason about, and provides us cross-DC replication and topology-awareness. It took years to get to this point though, and to build the operational expertise required to run Cassandra. Only recently have we gotten to the point where teams are able to self-service provision their own Cassandra clusters.

This is a resilient, scalable solution, but if I were to quit tomorrow and start a five-person startup, there is no way I would consider C* as a workable solution.