Benefiting from the substantial improvement of wireless sensor networks and embedded computing devices, the Industrial Internet of Things (IIoT), which is dedicated to alleviating computation barriers caused by hardware limitations, has been recommended for widespread deployment. In addition, as a decentralized collaboration mechanism, micro-services are being converted to be the preferred “collaborative” paradigm to enhance the working and management efficiency of the IIoT. However, IIoT applications, regarded as a primary generator of data, tend to be large-scale and time-sensitive, thus posing considerable challenges to IIoT networks in terms of improving quality-of-service (QoS) and the scheduling of micro-services. In view of these factors, this study attempts to construct a cloud-edge collaborative network architecture and to formulate total time consumption, resource utilization, and fairness as a mathematical model, while employing an offsite task placement mechanism to protect users' privacy. To this end, this study devised a QoS-aware microservice scheduling method named QCEMS. The results of extensive comparative experiments on two mutually exclusive scenarios with three different contrasting methods demonstrate that as the scale of micro-services or the number of containers increases, the proposed method can perform comparatively well on three pre-defined objectives (i.e., latency, resource utilization, and fairness) under privacy constraints.