I used it with EBS volumes - mongodb datadir and also for rabbitmq datadir - works wonderful. If a pod fails it detaches then comes right back up within a few minutes.

We only have a single mongodb and rabbitmq pod since they aren't mission critical if they go down. We had the mongodb host fail and by the time I got paged and woke up the OK page came since kubernetes did its job and brought it back online