A single point of failure is an advantage?
As a king, what would you prefer - to defend against one big kingdom that wants all of your land, or to defend against three big kingdoms that each want a third of your land?
When number of points of failure trade off against your attack surface, you can't just say "more PoFs is better".
If we were to use that analogy, a more accurate metric would be either having 1 large centralized kingdom or 100 small, autonomous, independent kingdoms that each have fundamentally different weaknesses. If one fails, you still have 99.
But even that fails to understand the dynamic of biological systems. There's mutually beneficial, symbiotic relationships between crops and many creatures (bees, worms, etc).
Presenting it as always in conflict is inaccurate.
The fundamentally different approach is to not look at pest solutions through a genocidal chemical warfare lens but instead to figure out how to maintain a sustainable relationship and have all of nature working for you instead of setting most of it as against you.
In context of sustainable ecosystems, I agree with your detailed explanation. Controlling and maintaining a complex system of feedback loops (that we still don't completely understand) is a very difficult task though, so I'm not surprised that currently the most crude methods (genocidal chemical warfare) are also the cheapest / most efficient.
I wonder if this isn't a good field to benefit from computational models though - from tools to design, predict and manage natural feedback loops in agriculture. I don't know, maybe it's already a thing in the industry?
Blanket chemical warfare is like feedback linearization. Wasteful in terms of energy and has other flaws associated with stability. There is a better approach, e.g., passivity based control, which requires understanding the underlying nonlinear dynamical system.
Interestingly this has been the same problem in medicine. However, there are more nuanced control-oriented medical techniques being developed (need better observers and actuating mechanisms).