It might be my ignorance of the topic but it seems to me that behavioral economics seems to have a major blindspot with respect to the computational & informational complexity costs of decision making processes (which are as real as any other kind of cost).
This could also be because it's more "interesting" to come to conclusions like "here's a pernicious irrational bias that most people have" instead of "people are essentially rational within their informational and computational limits".
Also robustness costs, -- many heuristics that get poked at by champions of behavioral economics are behaviors that tend to give improved results when reasoning with noisy inputs (and, in particular, inputs with unknown levels of noise or even malicious distortions) compared to a formal decision theory that doesn't handle low quality inputs well.
These limitations are only a concern when people attempt to suggest that behavioral economics is telling us to rework how we live our lives. Generally such claims are unjustified extrapolations from very basic studies which in no way support the pop-economics advice. Like if someone did some experiments and determined that submersing people in water often killed the test subjects, then other people ran to the presses recommending that no one drink anymore because water was proven to kill in scientific studies... :)
I'm not sure why the focus here is on behavioural economics specifically. Ignoring decision making complexity is really the domain of traditional microeconomics. If anything, behavioural economics is the path by which economics as a whole has the best chance of digging itself out of the hole it is in, because it at least acknowledges that you have to look at humans a little bit rather than just inventing theories from whole cloth.
I think you're correct in your assessment, but there's more to it. Specifically we must remember that evolution finds more useful solutions over time, but not necessarily the most useful solutions. A local minima of cost can be found while ignoring an even lower point of cost on the other side of a hill. It would be ridiculous to expect us to have found the most rational course of action in a given situation, even accounting for the mental costs you mentioned.
Complexity costs of decision making is actually at the core of behavior economics, for instance one of the biggest insights (which sounds really obvious in hindsight) is that people stick to the defaults. You can see this insight everywhere from opt-out dark patterns and default opt-in organ donor systems but according to classical economics something being default opt-in or default opt-out shouldn't matter because people will take the time to make a most optimal decision possible with all possible information.
Most of the talk of 'rational' vs 'irrational' is because it is in opposition to the classical economics assumptions that people act perfectly rational with perfect information. The more full explanation a behavioral economist would probably give you would be something about how the human mind is optimized for a very different environment and takes short cuts appropriate to that environment that are not helpful in the current environment.
This could also be because it's more "interesting" to come to conclusions like "here's a pernicious irrational bias that most people have" instead of "people are essentially rational within their informational and computational limits".