Improving observability in Event Sourcing systems

Abstract

Abstract Event Sourcing (ES) systems use an event log with the double purpose of keeping application state and providing decoupled communication. While ES systems keep track of all business events, other untracked events, either from internal components or from external systems may still cause failures. Determining the root cause of such failures usually involves complex procedures based on replaying the event log. Unlike this, in distributed systems, developers often instrument the source code, for the sake of improving observability and perform tracing on workflows and data. Adding tracing to ES thus seems like an unexplored and powerful approach to improve the observability of the system. In this paper, we suggest possible implementations of the idea and discuss their merits. These include the adoption of well-known tracing-related tools and standards in ES systems, with the respective advantages for root-cause analysis, anomaly detection, profiling and others.