I am having a hard time wrapping my head around an apparent contradiction between the success of life and the compounded sources of error in living processes.  Lifeforms have evolved complicated systems for development, replication, metabolism, etc.  In one respect, the opportunity for errors to pile up to a critical level seems like a statistical certainty...to the point that nothing very complicated should be able to evolve.  In terms of information theory, life is gobbling up sunlight (information, potential energy) and able to store or utilize it to perfrom some simple operation along its path to entropy.  But there are so many life forms and systems that are not simple; they rely on thousands of layered, codependent operations.  Each of those procesesses is not only inefficient from a physics perspective but also laden with some error rate.  How does life beat the compounding of such error rates? 
     To make my question a little more tangible,  consider human development to birth.  Evodevo has shined light on the amazing cascading effects of genetic switches that require particular timing and environmental cues as the zygote develops.  How many critical benchmarks are there in the path of development, how many opportunities for error?  Even if 40% of fertilized eggs miscarry, this still leaves 60% success completion of a mind-boggling number of steps that had to occur correctly. 
     I think of this sometimes while driving my daily commute in Houston.  As a firefigter, I've responded to many occaisions where some critical level of error in the complex system of driver interactions has ended in catastophe.  However, auto-accidents statistically represent only the tiniest hiccup in a system which seems to maintain an impossible level of success for an interaction that is so frequent, complicated, and dangerous.  I extend the metaphor to thinking of myself just walking around...why not a sudden aneurysm, lethal bacterial infection, some metabolic run-away event, something that causes me to leak out, explode, or implode.  Any one of those events may not be frequent, but add all the possibilites up and it seems absolutely astounding that body continues to win day after day for an average of 80 or so years...
     What concepts am I missing that rationalize the success of life on the backdrop of entropy?
     Thanks so much in advance!

its a great question and there is no easy answer I'm afraid.

The variable missing from your post is the hundreds of millions to billions of years of evolution that have finely honed the various pathways you describe in a system.

At the start of every unoptimised system there will be a point where it will still work a bit and give a read out. Tweak 5 bits of the system each by say 2-5% it works a whole heap better. make that hundreds or thousands of components in a complex biological system and optimisations over hundreds of millions of years then it works to close to a peak of optimisation.

Then as you say there are constant mutations that decrease the effectiveness a bit and occasionally are so catastrophic that the pathway stops or drops below a minimum level of effectiveness. That leads to disease, foetal malformation or death.

To put it another way in massively well optimised system there is enough headroom to tolerate some degree of failure - most of the time.