that board looks like the cheap crap and likely the components are much more durable then the PCB. And again if you lift the components they get more convection cooling and run at a lower temperature.
I replaced a zener diode with a bigger one in my welding machine (igbt driver supply) and the bigger diode did not make much of a difference on a thermal camera test. Ceramic made a bit of a difference compared to plastic IIRC. The board was also damaged there... but I am talking small difference I think like less then 10C, and it was ugly and scrunched together. I don't think there is component magic you can work, it needs either heat sink or elevation and further then that, some kind of forced air if absolutely required. I wanna say the difference between a 900mW? and a 1.5 watt zener diode was 5 degrees C on the thermal camera.
And also, if there is a catastrophic failure, the elevation will probobly prevent grevious board damage with the lead bend.
Also, the diode lasted approx a few seconds longer (something was messed up causing them to bust, I replaced the diode like 10 times trouble shooting the PCB lol. A bigger resistor is impressively more tough, it seems that a bigger diode.. not so much. It seems that when zeners fail, they fail quick, regardless if a slightly higher power is substituted. I figure a 5 watt diode might take the condition a little longer but in the end, that might just be a bigger glowing turd on the board doing more damage. I don't renember exactly, but I think I even tried to put a 2 or 3 watt diode in there, and it still got smoked mad fast. I just wanted more time to see if I could find something that was heating up. With a resistor you can sometimes put a beast in place temporarily to let the circuit run for a bit to see what is over heating without just shorting it out or whatever, at the risk of blowing up a source.... but in some cases you have spare parts for a simple easy to access source when the trouble is hidden in a mess of complicated bullshit

. So long you don't damage a transformer, replacing some power transistor after you are done.. who gives a shit? (especially if you are worried about sequencing problems or whatever to prevent you from easily running a power supply into that node)
Maybe I am wrong, but empirically I noticed that, the 'overload' resistance of a diode is meager in relation to size compared with a resistor, in typical circuit conditions. IDK if its the heat being generated mostly in the junction (spot heating) or what, but it seems to just not have much more strength, while a power resistor gets more durability scaling.