it will be less and less rewarding to keep adding hashing power for the miners to protect the network and more and more tempting to try a 51% attack even if it just for breaking the confidence in the network.
This is a pretty weak assumption. We don't make any calls for necessity in the Bitcoin hashrate growing, there's no security assumption around that, it's just an observed property that presently it happens to be growing.
Bitcoin miners are assumed to be operating with a single goal: profit. If somehow causing the thing which is giving you profit to fail is adventurous to you, then we can just rest easily on the knowledge that the underlying basis for Bitcoin to exist is incorrect.
The incentive to play the game fair and keep mining is pretty high, but what will happen once that reward get very low like 1 bitcoin or even less, even if the price get big like > 100 000 dollar it will be less and less rewarding
If the cost for operating the hashrate exceeds the revenue generated, miners don't mine. The whole system ends up with an equilibrium of having barely profitable mining hardware- regardless of how much profit there is to be made doing it. If the fee volumes don't ultimately cover for this, there will be less mining.
forever keep a sufficient amount of machines ready to be powered ON at any sign of a 51% attack attempt
You assume incorrectly. There is not large farms of idle hardware simply laying around. Large facilities require a significant amount of upkeep, large amounts of power are not available instantaneously, and if there was large amounts of ready-to-go obsolete hardware, it would doubtlessly be so inefficient as to be worthless towards any cause.
so what could be the solution to keep the network protected
Failure of the incentives means complete failure of Bitcoin as a whole, to pretend otherwise, or to pretend that there's some secret bullet which hasn't already been deployed, is disingenuous.