Like I said in the forum history, I should have paid for persistent storage, but this data lose could have been avoided during a planned reboot.

FlyIO knew that there was no storage attached to the redis instance and there would be data loss, but they still chose to reboot it ahead of schedule.