The entitlement is quite staggering, as evidenced by the comment you replied to. This is like Cocal Cola putting a free refill station on the street and a bunch of people rolling up with tanker trucks, then getting upset when it's taken away.

Given all the talk about how scalable some of these services are, they end up being pretty easy to break. I'm not on data hoarders, but I personally "broke" my crashplan account a year or two ago. All it took was having 3TB of data, mostly a bunch of family pictures, some home movies, a half dozen cloned git repos with a crapton of small .c files in them and having an account for >5 years. My client got wedged in some kind of server reindex operation (or whatever was happening on the server side in the datacenter) that took longer than the 24 hour window to complete. After they told me to reduce my retention window because I had to many "objects" it mysteriously fixed itself a week or two later after basically being offline for a couple months.

Similarly some software I wrote a couple years ago, that wrote a bunch of small files into a single S3 bucket caused a lot of grief at a 3rd party S3 like service because they never imagined someone would write a couple TB of data into a single bucket (actually it used multiple buckets but each one was a TB or two of <1k files).