In a microservice architecture, solutions are built through collaboration of distributed services across networks. In the Oil & Gas industry, in exploration and production phases, organization units executes different services over several diverse datasets. Geological data usually is in high volume and encompasses different kinds of data objects, with diverse structure and nature, such as seismic data, seismic horizon and well data. Querying, processing, and composing geological data presents strong demands for domain knowledge representation and reasoning, and tailored processing techniques. This work presents an application of microservices architecture and polyglot persistence technologies to handle the requirements of geological data in the Oil & Gas domain. This architecture allows parties communicate in a light way, while encapsulating processing and data access to a geological database and to a knowledge base. It also works as a common layer, composing parties' services results for data consumption by clients. We exemplify the proposal by presenting and analyzing its use in a real scenario which includes some of the implemented queries in a developed system to support geological data analysis. We present the main characteristics of the system and highlights lessons learned in its development.