HP OMEN 15 and 17 owners lounge

Your GPU temps were always that high? 2060 shouldn't run that hot, but (again, I don't know your model) in the 2019 models HP used two different fansinks depending on the tier CPU/GPU combo, dividing line being below i7/2070. If you have a weaker fansink you're stuck with the other techniques to make the best of it.
Both MX4 and Kryonaut worked extremely well on my 2080. The Kryonaut gave a few degrees cooler GPU temps and what's currently on it. The Nvidia die is a huge contact patch so honestly any quality paste will be better than dried out OEM, it's that skinny Intel one that's really tricky.
If it's your first time, download and have a read of the service manual for your model to familiarize yourself. HP is one of the few manufacturers that makes it available to us. It's easiest to find via support, use your product number, manuals, Maintenance and Service Guide. HP uses soft screws so make sure you have a decent toolkit.

 

Mine should be a 2018. For reference, i found this guide https://it.ifixit.com/Guida/HP+OMEN++15-dc0030nr+Thermal+Paste+Replacement/112630
It's not the exact same laptop (different code, very likely to have different configuration) but visually everything is the same

 

Looks good, if you're up for it. You have a warmer GPU and cooler CPU than me because that design has a dedicated CPU heat pipe, the 2019 models don't (both are shared with GPU). You are certainly fine just waiting to see if it gets hotter over time, as long are your GPU isn't throttling and slowing your games (use Afterburner to monitor).
If you do tackle it, just be mindful that the two halves are held together by only that one heatpipe--don't stress or bend it. The very first thing to do as you remove it is flip it over and account for all the pads--in case they ripped, got caught, assess if you can reuse them. It's generally fine if they tear in half, stick on the chips, or stick to the heatsink. It's not fine if they get lost or irrecoverably mangled. Replacing them as well will make the project more complicated.

 

I seem to be lucky then, I don't have any issues with a 28 second throttle timer; just have Omen set to performance mode (auto fans); UV setup, and maximum power set via TS to 100W; CPU never seems to go above ~89W even when fully loaded, and doesn't seem to throttle due to power under a long load.

It WAS getting rather toasty playing certain games or running certain loads though that do tax 8+ threads though so I literally in the last couple of days turned the all core turbo down by 2 bins (rest are left as standard). Knocks me from 4GHz all core to 3.8GHz when all 16 threads are loaded but a rather substantial 10 degrees plus off the CPU temps under all core load, whilst still turboing as standard when less cores/threads are in use.

Given the shared heatsink on the CPU/GPU and that my GPU is running OC'd I figured that 1/2 bins was a reasonable compromise; only losing a few percent of total CPU power but gaining a considerable extra amount of room away from the throttle temps .

Still toasty but mid to late 80s maximum temp detected running a TS bench 960M load across the 16 threads down from 98 degrees max temp is a noticeable improvement, and should enable the system to run longer and maintain peak clocks longer with no/fewer throttle spikes or similar, especially with weather due to get warmer over the next few months. Given how Nvidia's turbo clocks work as well; dropping those degrees off the heatsink load may even help the GPU maintain a slightly higher clock average when I'm not hitting the power cap also.
(Albeit I'm maybe not helping myself there as I have the VRAM overclocked as well).

 

Similar to what I did, but for gaming I did all the core multiples to 38. The problem in this machine when the GPU already has the pipes at 75 C isn't 8 cores going to 4.1, it's 1 or 2 cores going up to 4.7. That's a very small hotspot, and the temp is read as the hottest core. That's nowhere near a 90W load, and I haven't seen any game hit even half that.
I'm a little annoyed that I just bought a 4K monitor for more real estate. With Gsync, the 3080 would idle at 300 to drive the 1080p I had been using with base temps around 39, now it's 1000 MHz and 30 something watts, and 47. That gives me even less CPU headroom. I'll try some demanding games and see what it costs me. Worst comes to worst I can turn the monitor off while gaming, although that's where I keep afterburner graphs.