Pure speculation:
- Seals had a component failure, and the replacement component was also defective. It would not be readily apparent. They would change many other things before they realized it
- A power surge caused multiple, simultaneous equipment failures
- There was an unknown change to their internet connection parameters
- They had a recent change in a key tech support position. Loss of corporate knowledge
- They introduced a bad cable or connector during the recovery
As long as we're speculating wildly, I was thinking that maybe something broke on the old server and someone made a snap call that it would be easier (if perhaps not faster) to push out the new version than put another band aid on the old server/software version.