Assumption of thermal equilibrium for ensembles

[Coffee_]

Take for example a canonical ensemble, to use the derived distributions and probabilities, the considered system has to be in thermal equilibrium with the heat bath.

Where in the derivation of the probabilities and distributions do we usually assume this? There is a point that I can identify assuming a constant temperature for the heat bath yes, but I can't find where we assume that the temperature of the system is equal to the temperature of the heat bath.

 

[MarcusAgrippa]

You are deducing the results for a system in thermal interaction with an heat bath by considering the heat bath + the system as an isolated system and using the max-entropy principle for the combined system. See H B Callen's book on thermodynamics.