If the glideslope was unreliable why was it not taken out of service and a NOTAM issued.
Leaving a faulty ILS in service to trap the unwary is dangerous bad practice, someone deserves some serious discipline to be applied.

In the U.S. they do where TRACON or en route radar is available: MSAW and EMSAW. This was an issue in the KAL 801 747 CFIT at Guam, because ATC had the MSAW disabled.

Only RNP AR. Requires special quals. Here are the four charts for 7R:
https://cimg8.ibsrv.net/gimg/pprune....1dd79ff5fa.jpg
https://cimg1.ibsrv.net/gimg/pprune....dae6ad9dbb.jpg
https://cimg0.ibsrv.net/gimg/pprune....f318a89451.jpg

Well the autopilot doesn't think it just follows the signal whether it is correct of false... A pilot who has some common sense (and it seems a lot of them don't), who is tracking a G/S with a regular rate of descent on a stable approach will not all of a sudden, if he is competent, simply push the nose hard over to follow a G/S towards the sea especially if he was warned twice by ATC prior to the approach that there might be some problems with the G/S.

Puzzle me this....

TEM: the threat is the dodgy GS signal, broadcast on ATIS and further backed up by ATC to the crew. How should a crew mitigate this threat to avoid making an error? Fly the LOC only approach (the weather was more than adequate) or if allowed an RNAV approach. It would appear that TEM was not effectively applied by this crew....

Oh please. FMC has track distance to the airport. From once it's loaded, even if it's 16 hrs away. It's simple math, 300 x track distance + altitude. It's so simple people can even do decimals for a quick estimate. In this case 2.6 = 2 x 300' = 600' .6 x 300' (standby, left me find a calculator...) = 1800, drop the zero, 180. Now you need to do something with the 180. Well 3 miles is 900', and 2 miles is 600', so do you think you add or subtract the 180' to the 600'? Most people would add that. So 600' + 180' = 780'. Report says they were at 200' so they were approx. 580' below glide slope.

Reading your post and I'm starting to realize why guys have the speed brakes out diving to 2000' 25 nm from the runway. And yes, after leveling off intercepted a false glide slope and didn't realize it. Where is this DTG function on the FMC? PROG page 1/4.

Here's the really sad part - they were VFR (2214Z) -

VHHH 192230Z 09012KT 9999 FEW022 SCT035 24/19 Q1019 NOSIG=
VHHH 192200Z 09010KT 9999 FEW020 SCT030 24/19 Q1018 NOSIG=

I would have thought they were flying IFR but in VMC.

Does anyone look out of the window ?
If the approach is correct the touchdown point will always be at the same point on the windscreen. If it in the wrong place it is time to worry.

There were minimum 2 in the cockpit I assume. .In my (old) days the PF was looking down at the ILS bars,scanning speed/ altitude the PNF was looking outside.and cross checking parameters.
looking at recent accidents and incidents like this one it would seem that many today are all looking at the bars and nobody is scanning the other parameters.or bother to look outside before minimum even when VMC.
Is is how the new training is made?. Full trust on automation ?

A fluctuating signal can bring the aircraft into a bad situation really fast. I’m sorry, but I’m one of those lazy pilots who tend to be a bit relaxed when I fly. I don’t sit there and read out mile vs altitude every single mile when I do an ILS.
A sudden GS pitch down on short approach followed by GPWS and go around. Good call.
The GS should not have been on the air.

I suppose if one was determined to do the ILS but not wanting the aircraft to be following any fluctuations(happens in Seattle as well), they could just use the V/S function while following the glideslope in general. Especially with the weather conditions being beautiful VMC, they could just be mostly visual. Back it all up by confirming 1600 feet when passing VH721, continuing at about 7-800 fpm, checking distance to go versus altitude continues to make sense along with the sight picture out the window.

punkaolouver finally nailed it.

This is exactly what we do as regular operators into HKG on 07R. Happens frequently. We always shoot the ILS, aware of the possibility of GS fluctuation. The RNAV is available, but we only bother if it's low cloud or vis. If the GS plays up (about 1 in 10 times), just select VS 7-800 fpm or disconnect and hand fly. Easy peasy.

However, this is common sense, experience and airmanship we're talking about here... Sadly not something prevalent in many operators in this neck of the woods, especially the airline were discussing or the muppets to the north of us.

VHHH 192230Z 09012KT 9999 FEW022 SCT035 24/19 Q1019 NOSIG=

Was the PAPI u/s?

The overall pattern here is the habitual preference for technoiogy over nature.

couple of months ago, was also established on the ILS 7R, autopilot engaged, stable and all of the sudden aircraft pitched down very rapidly and violently.....immediately disconnected ap, recycled FDs, and able to salvage approach.
Yes they offered Rnav on the Atis but company not authorized for Rnav, and believe Atis said possible GS oscillations- thought they meant it might scallop a little up and down, really wasn't expecting that.