My guess is that these are standard HF ballasts and T8 lamps running at an incorrect voltage, it might be under voltage due to some fault or undersized generator. There are videos on Youtube showing this effect in Walmart stores when the power goes out.

A similar effect although not as bad as this can happen if you replace a remote ballast MH fixture with HF fluorescent without disconnecting the remote ballast first. It happened at my place of work, but it is a long story.

As Sol mentionned, this looks like an instability caused by high impedance feed to the ballast. I would guess for a loose Neutral somewhere...

I'm guessed that this is an open neutral case, as even preheat will flash like mad in an open neutral case.
What wonders me, is how the ballasts survives this, if they are electronic. LED would blow instantly in this case.

Neither the electronic, nor the LED's would blow instantly.
The semiconductors are usually rated for at least 600Vpeak on the mains (with decent makers), where the 400Vrms is still just below (there is not much margin, but by far not all the voltage would be across a single ballast).
The electrolytics are rated either for 400V (NPF) or 450V (power factor corrected), what mean a short term overvoltage capability of at least 600V (with quality makers), so again no sudden death. Of course, the capacitors will degrade way quicker than usual, but these faults do not last for as long.
And if the LED's are electrolytic free (as many, mainly new designs are) and are made using decently rated components, they will survive it for very long time (the reduced margin for an overvoltage would make it more prone to other disturbances, so it would be their statistical occurence, what will kill the LED).
And ironically the electronic won't get full overvoltage, when only those are behind the Neutral fault, as the mains won't be able to feed the 3'rd harmonics to any of the phase (this component sum up in phase on the Neutral and without it it won't keep the peak voltage; all the time with NPF, during the overvoltage with HPF as well; the overvoltage will cause the PFC effectively to de-engage itself)

Magnetic will have way bigger problem, when this last for longer time: The overvoltage will overheat the chokes, as they are usually designed thermally, as well as magnetically, with nearly no margin at all. So even a small overvoltage can kill them, if it is present for too long... Well, the mass is the only thing, what protect the magnetic from an instant death, but that work only for short time...