This article focuses on the issues of fault-tolerance and scalability of .NET-based architectures with a microservices design pattern. Message queues are one of the most widely used technologies for building fault-tolerant distributed systems. The motivation for conducting this study lies in the need to increase the reliability and flexibility of modern software systems, which are becoming increasingly distributed and complex. The starting point of the analysis is the limitations of synchronous, request/response communication between microservices, tight temporal coupling, increased latency and fragility under load, and a heightened risk of cascading failures. The article demonstrates how the adoption of message brokers and asynchronous messaging patterns mitigates these problems and provides a basis for building more resilient and scalable systems. The goal of this study is to gather and systematize queue-based approaches to implementing fault tolerance and scalability in microservices in the .NET ecosystem. The work undertakes a comprehensive analysis of architectural decisions, encompassing a comparison of Apache Kafka and RabbitMQ technologies; an examination of the Saga, CQRS, Circuit Breaker, and Retry patterns; and an evaluation of quantitative efficiency metrics for event-driven approaches (EDA). The scientific novelty lies in an integrated synthesis of the technological, architectural, and analytical aspects of microservice system design. The main conclusion is that a message broker should be chosen based on load and the business case. Kafka is preferred if the workload requires high-performance streaming and data transfer into analytical data pipelines. RabbitMQ is preferred for complex routing or background job processing. Patterns like Circuit Breakers and Retry with Exponential Backoff can be used to protect against cascading failures. Pattern-based approaches such as the Saga and CQRS patterns provide eventual consistency in distributed systems. The article is aimed at software architects, senior developers and researchers who design scalable, fault tolerant microservice systems on the Microsoft .NET platform.