What is the definition of a clock in relativity theory?
an object that measures the rate of time flow.
The quest for a good clock has developed slowly and empirically over time. A rather interesting show on the topic is "Longitude", in which the first naval chronometer was developed. The goal wa simple - to develop a clock that could be taken to sea, and still keep good enough time so that a person could determine their latitude by measuring the time at which the sun reached it's high point in the sky.
As a practical matter, we do have a definition for a "good clock" nowadays. But it's not a theoretical definition, it's one from a standards comission. I refer, of course to the SI second. And, to tie it in with my pervious comment, our "good clock" nowadays is also very small - it's an atom.
It is a good thing, in my opinion, that we define our clocks empirically (through standards comittee's) rather than through theory. Defining a clock theoretically is not productive, as science has to be based ultimately on observation.
A friend of mine is a GR expert. He wrote an article on the nature of time and what is meant by a "clock". Its a great article and a good read. See
Separate names with a comma.