Distributed Tracing & Centralised Log Management

This project aims to showcase 2 key features of a microservices application. The first is distributed tracing. The second is the implementation of a centralised log management system. This application has been developed using Spring Boot

Spring Cloud Sleuth & Zipkin

I am using Spring Cloud Sleuth for distributed tracing and Zipkin for visualisation.

Logstash, Elasticsearch & Kibana

Spring is logging using Slf4j with Logback configured to write to Logstash. Logstash is writing to Elasticsearch. Kinana is used to visualise the logs

Running the application

Follow these steps:

1. Build the Spring application

> ./mvnw install

1. Start the infrastructure

> docker-compose up

1. Start service 1

> ./mvnw spring-boot:run -Dspring.application.name=upstream -Dserver.port=8081 -Ddownstream.server.url=http://localhost:8082

1. Start service 2

> ./mvnw spring-boot:run -Dspring.application.name=middle -Dserver.port=8082 -Ddownstream.server.url=http://localhost:8083

1. Start service 3

> ./mvnw spring-boot:run -Dspring.application.name=downstream -Dserver.port=8083

Invoking the endpoints

1. Call the rest endpoing a few times.

> curl localhost:8081/zipkin
upstream -> middle -> downstream

Configure Kibana and view the logs

1. Open Kibana http://localhost:5601/
2. Under "Management" create an Index Pattern. logstash-* and @timestamp
3. Choose "Discover" to view the log entries

NB. If you want to filter on a particular trace you can use something similar to this

message: ">>> Call*" AND traceId: "7f3a243d1b53720f"

View the tracing in Zipkin

1. Open Zipkin http://localhost:9411/
2. Choose "Find Traces"