Tracing with Spring Cloud Sleuth, OpenTelemetry and Grafana

This code and content is based on [https://github.com/thombergs/code-examples/tree/master/spring-cloud/tracing], with its great companion article Tracing with Spring Boot, OpenTelemetry, and Jaeger

Running and exploring this demo

• run ./mvnw clean install to build the two Spring Boot applications (api-service and customer-service)
• create the dedicated docker network so that services can reach each other and grafana with docker network create observability_tutorial
• run docker compose -f grafana/docker-compose.yml up -d to start grafana in the background
• run the services with docker compose up --build
• call http://localhost:8080/customers/<ID> (where ID is a number from 1 to 50)
• open http://localhost:3000/explore in you browser while watching for the logs of the api-service.
• Notice the trace ids in the logs, second value into brackets. In grafana, go to the Explore View, select Tempo in the top left drop-down menu, and enter a Trace ID to see its content.

The above HTTP call goes to the api-service, which will call the customer-service for additional information. This will create a trace across both services, as should be evident in the logs with the same trace id.

The docker compose command also starts up an OpenTelemetry Collector, to which the Spring Boot apps send their traces. The OpenTelemetry Collector, in turn, sends the traces to Grafana.

Switching to Kafka demo

• run git switch kafka
• recompile with ./mvnw clean install
• run docker compose -f kafka/docker-compose.yml up -d to start kafka in the background
• run the services with docker compose up --build
• start back with the same example