Microservice architecture is a design pattern where a business application comprises individual, functional services that can scale independently, offering flexibility and reduced risk when implementing changes. This architecture contrasts with the monolithic design, which traditionally utilizes a single database for the entire application. In microservices, each service often has its database, but this decentralized model introduces challenges, particularly in managing queries and transactions across services. To address these, various database management patterns are used, such as the database-per-service pattern, Saga pattern, API composition pattern, and event sourcing pattern. The article explores two prominent data storage approaches for microservices: polyglot persistence and multi-model databases. Polyglot persistence allows each microservice to utilize a specialized database suited to its needs, supporting diverse data models like relational, graph, or document-based. Meanwhile, multi-model databases provide a single engine capable of handling multiple data types, offering ease of management and a unified semantic model. The choice between these approaches depends on factors like enterprise size, team expertise, and specific use case requirements. Polyglot persistence is recommended for large enterprises with dedicated database teams, while multi-model databases can benefit startups looking for simplified management. As microservices gain popularity, understanding these patterns and storage solutions becomes vital for optimizing development processes.