So I have a convoy S6, with a solid pill and an xml2 on copper noctigon running at 3.04amps.
Should I be worried about the LED failing after running on high for a long time?
The body gets very warm to hold, but if I’m running on a 3400mah panasonic, giving me a likely runtime about 60min. should I be able to leave the light on high and not need to worry about things frying?
Is 60min. too much and 30min. more realistic?
Or should I be expecting 5min. max?
If you can still touch the light without feeling like you're getting burned, then the LED will be fine on a Noctigon. The light might be able to overheat if left on high unattended, but it's unlikely if you're holding the light. The light will probably hit maximum heat saturation after about 5 minutes.
Your hand is a great heat-sink for flashlights.
5 Minutes does sound like a good number. Does a gloved hand help heat-sinking? 
Now say I leave the light on unattended, should I expect de-solder first?
If you're tailstanding, you could drop it in a glass of water...that will limit the temperature on the outside of the light to 212°F.
Led emitters are tested/rated at 85 degree Celcius (if Cree) and that is too hot to hold for anyone I know. In my experience dont worry. If the emitter gets really really hot it desolders itself or the wires to the star the emitter sits on. Overcurrent and voltage kills emitters. Not any temperature rise due to long running at high power.
Why should that not be a worry? If the + lead de-solders and contacts the grounded area all around it, the pill, the base of the mcpcb, tube wall, you've got a shorted cell, no?
Protected cells are highly recommended if you decide to thermal test them or make a habit of putting them in a risky situation.
After all, Li-Ion flashlights are just a sophisticated pipe bomb if things go horribly wrong.
To put things in perspective, solder melts at 400 deg F (200 deg C). This is twice boiling temperature.
Most likely the emitter will die before reaching this temperature. You driver becomes the current limit.
In the absolute worst condition (multiple failures), your driver is more likely to cause a dead short than the emitter wire (save breakage).
Simple solution, make sure your light can be held in your hands. Don’t overdrive your flashlight. Be mindful of what a flashlight is.
Cree uses very conservative numbers for their products. They are rated for continuous use in 85 deg C temps are their baseline.
This is already damn hot.
Gloves are an insulator so no, gloves are not good heat-sinks. The body has its own thermal regulator.
You draw the heat into your blood which is cycled throughout the body.
It is almost like having loosely coupled liquid cooling with a very large reservoir.
As a reference, I have done a tail stand on a 2.8A XM-L for 20 minutes before.
It was a C8 size light and it got too hot to hold, but not nearly hot enough to be worried about.
@ Gj: A sanyo/Panasonic or other quality IMR will not blow, Only die. Protected batteries does not even die.. Been there.
BTW: In my experience it is more likely that the emitter itself desolders and slides off the mcpcb. Of course it may still short.
@ Nightspy: Several members have reported that they have had this happen and the emitter is fine afterwards. So have I.
Dying is something to worry about! 
And a lot of people use pack pulls in these small lights.
Where's the emitter sliding to? Isn't it pretty much held in place by the reflector & centering ring? And if tail-standing, there's no force sliding it off sideways anyway.
My first LED reflow was on a hotplate and the dielectric layer slid off. :D
Been there, done that… it’s an over-rated pastime. J)
Nope, noctigon is the way to go, will it get hot sure, but that is what modes are for. 
This is why STAR firmware and the auto turbo timeout is so nice. 
Regardless of the perils.
For every member there are 20 that have bought that same light and run it on whatever craptastic battery they figured would be good/cheap/acessible. Very very few reports of flashlights blowing up.
And I maintain this: At 3 amps? You will be fine. And so will the light. It is something like 12 watts and the battery can only maintain that for precious few minutes. But if it is such a concern then by all means order one with 2,1 amps instead. But we all know you dont want to. Or you would have already.
I've run some 18650 zoomie flashlight for the full battery duration with a direct drive driver because I thought I needed the light. Inside a 4 story marine boiler during maintenance. Ambient was ~50 deg. C. When low batt warning kicked in I changed the battery and did it all over. I had to repair the refractory lining. For most of the time it was tailstanding. Some of the time it was clipped to my hearing protector and very close to my head. I could feel the heat. Though I did switch it to low on order not to be blinded. Buy good batteries. Even laptop pulls are good. Some of my best performers for daily driver lights are pink sanyo 2600 mAh from a HP laptop.
So maybe I am biased. Maybe I take chances. But no lithium battery is inherently safe. But then neither is NiMh. Or Led/acid. Or...
I can attest to what NightSpy says, last night I was doing some testing on my new light build (EA41_CK, an XP-L triple modded EA41 with big current boost). I’ve previously ran the light on max for well over 10 minutes carrying it around in my hand and it got very warm but not to hot to hold (I was even stepping the turbo back up after the timer). Last night I set the light down not thinking and within the 3 minute turbo the light go so hot it partially desoldered the GND lead causing a short to ground and the emitters to always stay lit (and that’s with the massive 6061 heatsink shoved down its throat and a new 32mm triple noctigon so thankfully the solder on the wire let go first and not the solder mounting the emitters so it was an easy fix)
Or have a thermal step down circuit and temperature sensor built into the light so that it only steps down if needed. That way if cooling is good the light maintains high output for longer periods or stays on maximum output mode. I believe that this is a solution used in some lights.
Thermal based step down (using a voltage divider made up of 2 thermoistors) is one of the many many functions of the Ultimate Flashlight Code as found on my BLF 17_dd_PIC drivers of anyone’s really interested (the UI does it but the current board doesn’t have thermosistor pad’s yet- revision coming).
It could also be done with a stand-alone board glued to the bottom of the emitter shelf to monitor LED / head temp instead of driver PCB temp.
It seems that if all the heat were coming from the LEDs, one of them would have either died or slid off the star. So perhaps the resistance of the wire was contributing to the heat.