Analyzing and Monitoring Kubernetes Microservices based on Distributed Tracing and Service Mesh

The microservice system architecture (MSA) outperforms the monolithic system architecture in terms of maintainability, extensibility, scalability, and fault tolerance. This is prompting a widescale migration of software systems from existing monolith systems to MSA. Most microservice systems utilize container technology for deployment. The fact that Kubernetes (K8s) provides a fully-fledged toolchain for managing container-based applications is prompting many organizations to adopt the K8s protocol for microservice system deployment and operations. Microservice monitoring is essential to the success of any service operation. The collection of logs and aggregation of metrics by most existing microservice monitoring systems is somewhat intrusive. Furthermore, the heterogeneity of Kubernetes technology means that most monitoring methods are inapplicable in situations where microservices are developed for a system using a variety of underlying languages and platforms. In the current study, we developed a monitoring mechanism that provides various metrics specific to microservice systems in a nonintrusive way. The proposed K8s-based microservice monitoring system, referred to as KMamiz (Kubernetes-based Microservice Analysis and Monitoring using Istio and Zipkin), enables the construction and visualization for service-level/endpoint-level dependency graphs and endpoint request chains, and the service cohesion/coupling analysis to enhance system quality for the development team.