There are many animal studies showing that moderate exercise causes extended health and many human epidemiological studies showing a robust correlation between moderate exercise, better health, and extended life expectancy. This should give us confidence in believing that yes, being sedentary is bad for us and the practice of at least moderate exercise is good for us. Showing causation in human studies is a real challenge, however, which is why there is still uncertainty over whether a lot of exercise is better, worse, or no different for longevity than the moderate 30 minutes a day recommended by most physicians. We can point to the fact that exceptional athletes such as Tour de France bicyclists live longer than the rest of the population, but is that because they exercise a great deal, or because only very robust individuals tend to become athletes competing at that level?
Separately, there is also the question of whether extremely high levels of exercise actually shorten life expectancy when practiced across a broader slice of the population, once you go on to consider more than just professional athletes. Again here we don't have much to go on in terms of causative relationships, but a brief tour of some of the relevant research is provided in this article:
All runners have heard about the tragedies. The marathoner Alberto Salazar, at the age of forty-eight, suffered a heart attack and lay dead for fourteen minutes before a stent opened up a blocked artery and saved his life. Hundreds of studies, as well as our own intuition, associate exercise with cardiac health. But, in recent years, a small group of cardiologists have advanced a hypothesis that suggests these tragedies may not be so shocking, after all: they believe that an excess of exercise actually damages the heart.
For those of us who believe that the "everything in moderation" rule applies to, well, everything, this argument makes sense. Exercise remains one of the best things you can do to improve your cardiovascular health, but you certainly do not need to run marathons to achieve the benefits. Moderate amounts of exercise throughout life are perfectly adequate. Athletes who exercise in extremes generally do so for reasons other than their health - competitiveness, professional requirement, compulsion. But recognizing that exercising more than a certain amount reaps no greater cardiovascular benefits is quite different than suggesting that this level of exercise causes cardiovascular harm.
After reviewing the data and interviewing experts in the field, my own impression is that among people without known cardiovascular disease there is no compelling data to suggest that mortality significantly differs between moderate and extreme exercisers. There is thus no way to precisely define an upper limit of exercise for an average healthy individual. I suspect, though, that part of what sustains the "too much exercise can kill you" myth is the widespread recognition of the so-called exercise paradox. That is, while consistent exercise decreases the likelihood that you will have a heart attack, if you are destined to have one it is more likely to happen while you are exercising. That's why no one can issue a blanket statement that extreme exercise is safe. It's also why so many researchers have attempted to figure out how to make extreme exercise as safe as possible.