The blue bottle experiment, and other similar reactions like the "chemical traffic light", are not infinitely repeatable. They are driven by the oxidization of glucose (or some other suitable reducing agent, such as ascorbic acid) by atmospheric oxygen, and will eventually stop working once the bottle runs out of either glucose or oxygen. However, each cycle only consumes a rather small amount of glucose and oxygen, so the reaction can be repeated quite many times.
Also, the blue indicator dye (methylene blue) itself is not used up in the reaction. Even if the reaction does "run out of fuel", it can be made to work again just by adding more glucose to the solution and/or simply by opening the bottle to let more oxygen in. However, since the glucose is not getting fully oxidized all the way to water and carbon dioxide in this reaction, the various incomplete oxidization products will gradually accumulate in the solution.
I'd expect that, if you kept adding more and more glucose to the solution, eventually you'd end up with a sticky sugary alkaline goo full of partly oxidized glucose products that was too viscous to be swirled around the flask any more. At that point, you'd really have no choice but to dispose of the experiment and start over with fresh ingredients.