Is temperature a microscopic or macroscopic concept?

It’s more of a macroscopic concept. The definition of the absolute temperature is T=dU/dS (taken at constant volume), where U is the total energy of some system and S is the entropy (a measure of how many quantum states are available to it) in conventional units. Now that derivative doesn’t make any sense for really microscopic systems, at least at any particular time. In fact, if you specify what quantum state a system is in (a reasonable thing to do for an atom, for example), S=0 by definition. So T is really a meaningful quantity only for macroscopic systems. However, for a microscopic system exchanging energy with the outside over time one can describe the probabilities of the system being in its different states. If those probabilities follow the Boltzmann distribution, then one can speak of the T of the microscopic system. The reason for them to follow that distribution, however, is the energy exchange with the macroscopic environment. So something macroscopic ends up involved either way.