This paper presents a theoretic approach to measurement of the throughput, respectively the bottleneck times of the production lines. The mathematical formalism utilized is characteristic for Discrete Event Systems approaches, and is about the Markov chains. Using this approach, we introduce definitions of bottlenecks, and we discuss their implications for production automation and preventative maintenance. The bottleneck of a production line is a machine that impedes the system performance in the strongest manner. The size of the buffer between the two machines is assumed to be finite. The approach is based on the sensitivity of the system production rate to machine reliability parameters.