The project WINNER aims to integrate and coordinate electromobility used through carsharing, the energy consumption of tenant households and the local production of electricity, e.g., by integrating photovoltaic systems into a smart local energy grid. While the various components correspond to the currently available standards, the integration has to be realised via a data processing and storage platform, the WINNER DataLab. The goal of this platform is to provide forecasts and optimisation plans to operate the WINNER setup as efficiently as possible. Each data processing component is encapsulated as a single service. We decided to use a microservice architecture and further an execution environment like container instances within a cloud infrastructure. This paper outlines the realisation as well as a report of our experiences while realising this project related microservice architecture. These experiences focus on development complexity, modifiability, testability, maintainability and scalability as well as dependencies and related lessons learned. Finally, we state, that the practical application of setups like this helps to concentrate on business models. It supports decoupling, helps in development to focus on the essential things and increases efficiency in operation, not least through good opportunities for scaling. However, it is required to mastering the complexity which currently requires clean planning, experience and a coordinated development process.