According to the report, the VDR stopped recording when the power went out. But, you state that the VDR has a battery backup. Why would it stop recording?
VDR stopped recording
sensor data. NTSB states that it continued recording audio data. The mics are built into (or plugged directly into) the VDR. The sensor data (helm angle, rudder angle, throttle command, throttle actual, heading, rate of turn, etc) are pulled from the Integrated Bridge System. Without looking at the particular installation on the Dali, I can tell you that the one's I've seen pull that info through an interconnection box (think router for the IBS) That interconnection box is probably powered by 120VAC (maybe 220) from the emergency switch board.
So the VDR itself has battery backup. The Electronic Charts and Radar displays, GPS, one of the VHF's, probably some other stuff that are all critical have battery backup, but some of the "2nd tier critical" stuff will be on the E. Switchboard and come back on when the E. Gen fires up. Remember that's supposed to be less than 45 secs, and there's only so many batteries you can shove under the console on a bridge.
The NTSB says that the VDR lost sensor date when the alarms went off (Audio alarms on all the battery powered stuff.) then picked up sensor data again a minute later when the power came back on. That sounds like power cycling to the interconnection box.
Look at this VDR system:
https://markogroup.com/en/ship-supply/voyage-data-recorder-rutter-vdr-nw6000/The mics are PoE directly to the Core unit, which has an Uninterruptable power supply, but a bunch of the sensors are going through the NW64XX Data Acquisition Unit, which is going to pull power from the E. Switch board. Some of the other sensors (RADARs and ECDIS mostly) may have to have digital-NMEA conversion boxes to pull the signal from the PC and make it readable to ship electronics depending on what exactly is installed. Those boxes may also be on E. Switchboard power.