The Art of Event-Driven Observability with OpenTelemetry

Ensuring the reliability of distributed systems requires observability of critical applications, and the ability to understand precisely what is happening in the environment. OpenTelemetry has standardized the way measurements are created and collected, leading many organizations to take advantage of tracing to understand their system. Building tracing in a microservices architecture is usually a straightforward concept where each service has dependencies, but when you’re using events or messages to exchange information between services, it’s another story. What should a distributed trace look like for EDA where you potentially have hundreds of asynchronous paths to a single message? Does it make sense to create an end-2-end trace for everything? Sometimes in event-driven architecture, the usage of span links will resolve our challenge by having several traces for a transaction, but all those traces will be attached with the help of span links. In this talk Rexed will explain the components of OpenTelemetry, share examples of useful traces from EDA, and talk about the purpose of span links in general and in context of EDA.