High availability in a microservice architecture

The purpose of this study was to investigate approaches to ensuring high availability of microservice systems with a focus on fault tolerance, scalability, and continuous operation of services. The study applied a comparative and analytical method to analyse technical solutions for ensuring high availability, systematise the characteristics of container orchestration platforms, and evaluate load balancing tools according to the criteria of performance, flexibility, reliability, and integration convenience. Fault-tolerance patterns – retry, circuit breaker, and fallback – that provide flexible error management, reduce the risk of cascading failures, and maintain system continuity are investigated. The study found that the behaviour of fault-tolerance patterns depends on the configuration of execution parameters, such as timeouts, retry limits, and conditions for activating fallback mechanisms. The effectiveness of such tools as NGINX, HAProxy, Envoy, and Amazon Web Services Elastic Load Balancing is evaluated in terms of their impact on the scalability and resilience of the architecture, as well as the possibility of automatic scaling on the example of Amazon Web Services and Google Cloud platforms. It was found that built-in autoscaling services ensure stable operation of services under variable load and enable a rapid response to peak loads. An overview of container orchestrators (Kubernetes, OpenShift, Amazon ES) was provided, among which Kubernetes is recognised as the most effective due to the support of self-healing mechanisms, distributed deployment, health checks, and integration with Continuous Integration/ Continuous Delivery. The findings of this study can serve as an analytical basis for designing sustainable microservice architectures in cloud and enterprise environments to improve the reliability, scalability, and continuity of business processes