{"id":17,"date":"2020-02-05T10:00:00","date_gmt":"2020-02-05T09:00:00","guid":{"rendered":"https:\/\/www.codemotion.com\/magazine\/tracing-performance-of-backend-services-written-on-java-with-sleuth-zipkin-elk-stack\/"},"modified":"2022-06-10T16:36:12","modified_gmt":"2022-06-10T14:36:12","slug":"tracing-performance-of-backend-services-written-on-java-with-sleuth-zipkin-elk-stack","status":"publish","type":"post","link":"https:\/\/www.codemotion.com\/magazine\/backend\/tracing-performance-of-backend-services-written-on-java-with-sleuth-zipkin-elk-stack\/","title":{"rendered":"Tracing performance of backend services written on Java with Sleuth, Zipkin & ELK stack"},"content":{"rendered":"\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\tTable Of Contents\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t
  1. Sleuth<\/a>
  2. Zipkin<\/a>
  3. ELK<\/a>
  4. Conclusion<\/a><\/ol>\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\n\n\n

    Logging<\/strong> and monitoring<\/strong> make up a very important part of any \u201chigh-load\u201d project<\/span> nowadays. The goal of this article is to provide an overview of what can be achieved in order to organise efficient logging and tracing of modern web applications<\/span> written on Java<\/span><\/strong>. A few tools configured to work together bringing great value to the project<\/span>.<\/p>\n\n\n\n

    In modern web applications<\/span>, especially if we talk about large enterprise projects, we can notice a very strong trend. It\u2019s called microservices<\/span><\/strong>. More and more huge monolithic applications<\/span> on the backend<\/span> side are being split into multiple small, reliable and more maintainable parts – so-called microservices. Taking this approach isn\u2019t an easy thing to do if you have to refactor a large project<\/span>, but it pays off ultimately: it\u2019s much easier to test, maintain, update and deploy<\/span> separate small applications<\/span>. That\u2019s why this approach is conquering the world of backend<\/span> today.<\/p>\n\n\n\n

    But with all those listed benefits, this way of designing applications<\/span> also brings potential problems and challenges that we have to consider. They\u2019re mostly related to the fact that the context of the whole application<\/span> is spread across the microservices. Because of that, it can be more complicated to debug<\/span> the application<\/span> and find out where exactly errors appear. There is quite a common way to handle errors efficiently and check performance<\/span> at the same time. It was the main topic of the talk \u201cTracing performance<\/span> of your service calls with some help of Sleuth, Zipkin & ELK<\/b>\u201d, given by Rafaela Breed<\/b> at Codemotion Amsterdam 2019<\/strong>.<\/p>\n\n\n\n

    To achieve clear visibility of what is going on in the whole application<\/span>, working on multiple microservices, first of all, we need to find a way to trace every request going through the chain of these microservices communicating to each other. And here the first tool in the list can help us.<\/p>\n\n\n\n

    Sleuth<\/h2>\n\n\n\n
    \"\"\/<\/figure><\/div>\n\n\n\n

    Sleuth<\/b> is a part of Java<\/span>\u2019s Spring Cloud<\/span> and it\u2019s heavily inspired by such tools as Dapper, Zipkin, and HTrace. Two main responsibilities of Sleuth that we\u2019re interested in are:<\/p>\n\n\n\n

    • It ensures that every request event initiated in a Spring application starts a new trace and creates a unique trace ID.<\/li>
    • It creates a span on every internal event, which also comes with a unique ID. It\u2019s being destroyed once the event is finished.<\/li><\/ul>\n\n\n\n
      \"\"\/<\/figure><\/div>\n\n\n\n

      As you can see from the diagram above, trace ID stays the same across the whole flow of the application<\/span>, but span IDs are different on each step.<\/p>\n\n\n\n

      So, let\u2019s see how we can add Sleuth into a project<\/span>. It\u2019s actually pretty straightforward, we just need to add it as a new dependency. In the case where the project<\/span> uses Maven<\/span> as a dependency manager (Gradle<\/span> is also supported), the code will look like this: