To swiftly and reliably monitor various information and the operational status of machinery across large smart manufacturing sites, this study introduces the use of LPWAN multi-mode network communication technology. This technology can automatically switch between LoRa and NB-IoT modes based on signal strength to ensure communication stability. Through wireless communication technologies such as LoRa and NB-IoT, the status information of machinery can be transmitted back to the cloud in real-time, facilitating user management. However, LPWAN multi-mode network communication modules often adopt a monolithic architecture, making maintenance and upgrades more difficult. When an application within the module needs upgrading, not only must the consistency between the development environment and the execution environment be ensured, but also a significant amount of time and resources must be spent on on-site deployment. With the development of cloud computing and virtualization technologies, containerized microservices architecture, which focuses on replacing functional modules with services, is set to become the mainstream for future industrial applications. Therefore, this study proposes a remote communication architecture based on container and microservices technologies. Utilizing the concept of microservices, this architecture divides LPWAN multi-mode network communication modules based on different functionalities and offers them to users in a more flexible service manner through containerization technology. This study also designs a mechanism to automate the entire service construction process, followed by the implementation of communication services’ automatic deployment through container management tools. Compared to manual deployment, this significantly reduces the waste of time and human resources. Finally, this study uses a large mobile pumping unit as a practical application case to verify the feasibility of the proposed architecture. In the context of flood prevention and disaster relief, large mobile pumping units are widely used to solve flooding issues. These pumps are often deployed in dangerous areas with poor signal reception, thereby also validating the value of the proposed architecture.