Motivated by Industry 4.0 applications, we consider quickest change detection (QCD) of an abrupt change in a process when its measurements are transmitted by a sensor over a lossy wireless link to a decision maker (DM). The sensor node samples measurements using a Bernoulli sampling process, and places the measurement samples in the transmit queue of its transmitter. The transmitter uses a retransmit-until-success transmission strategy to deliver packets to the DM over the lossy link, in which the packet losses are modeled as a Bernoulli process, with different loss probabilities before and after the change. We pose the QCD problem in the non-Bayesian setting under Lorden's framework, and propose a CUSUM algorithm. By defining a suitable Markov process, involving the DM measurements and the queue length process, we show that the problem reduces to QCD in a Markov process. Characterizing the information measure per measurement sample at the DM, we establish the asymptotic optimality of our algorithm when the false alarm rate tends to zero. Further, when the DM receives incomplete data due to channel loss, we present asymptotically optimal QCD algorithms by suitably modifying the CUSUM algorithm. We then explore the last-come-first-served (LCFS) queuing discipline at the sensor transmit queue to lower detection delay in the non-asymptotic case. Next, we consider the case of multiple sensors, each with its own wireless transmitter queue, and show that our analysis extends to the case of multiple homogeneous sensors. When the sensors are heterogeneous, we present a sensor scheduling algorithm that minimizes detection delay by balancing the trade-off between the age of the observations and their information content. Numerical analysis demonstrate trade-offs that can be used to optimize system design parameters in the non-asymptotic regime.