Method for ordered parallel event delivery in microservice systems using RabbitMQ, HTTP backup channels, and Saga pattern to increase viability and reduce MTTR

DOI: 10.31673/2412-9070.2026.318117

Abstract

The article investigates the problem of reliable, ordered, and fault-tolerant event delivery in microservice systems based on asynchronous message exchange. The relevance of the study is determined by the fact that, in distributed software systems, violation of event ordering, message duplication, or broker unavailability may lead to an inconsistent state of business processes, increased recovery time, and reduced system liveness. The purpose of the research is to develop a method for ordered parallel event delivery that combines the RabbitMQ message broker, a backup HTTP channel, and the Saga pattern for managing distributed transactions.
The introduction substantiates the need for an integrated approach to event delivery in microservice architecture, taking into account scalability, availability, and correctness requirements. The literature review analyzes existing approaches to event-driven architecture, message brokers, retry mechanisms, channel redundancy, and the Saga pattern. It is shown that most existing solutions consider these mechanisms separately, while their integration for simultaneous ordering, liveness assurance, and MTTR reduction remains insufficiently studied.
The methodological part proposes a system architecture that includes event producers, the RabbitMQ broker, microservice consumers, a Saga coordinator, an ordering control module, an event buffer, an idempotent processing mechanism, and a backup HTTP route. Each event is formalized as a structure containing a unique identifier, an ordering key, a sequence number, and a payload. For each ordering key, sequential event processing is guaranteed, while events with different keys can be processed in parallel. This approach makes it possible to combine high throughput with control over the logical order of execution.
The analytical modeling section presents the main indicators used to evaluate the effectiveness of the proposed method: successful delivery probability, ordering coefficient, mean time to recovery, and average delivery latency. The experimental validation was carried out in a Python 3.11 simulation environment under normal operation and broker failure scenarios. The obtained results show that the proposed method increases the probability of successful event delivery to 99.8–99.9%, provides an ordering coefficient of approximately 99.7%, and reduces MTTR from 957 ms to 263 ms, which corresponds to a reduction of more than 70%. At the same time, the increase in average delivery latency remains minor and amounts to approximately 3.7%.
The practical value of the results lies in the possibility of applying the proposed method to the development of enterprise and high-load microservice systems where guaranteed event delivery, preservation of event sequence, consistency of distributed transactions, and fast recovery after partial failures are critical requirements.

Keywords: microservice architecture, event-driven systems, RabbitMQ, Saga, backup HTTP channel, ordered event delivery, fault tolerance, MTTR.