[JNATION25] Let your Java application be truly observable! Let's dive into logs, traces and monitoring

Transcript

1. Let your Java application be truly observable! Let's dive into logs, traces and monitoring David PEQUEGNOT / Alexandre TOURET
2. Once upon a time a Friday afternoon… https://blog.worldline.tech/2020/01/22/back-to-basics-logging.html
3. 2024-06-10T10:44:54.444+02:00 ERROR 387865 --- [easypay-service] [o-auto-1-exec-5] o.a.c.c.C.[.[.[/].[dispatcherServlet] : Servlet.service() for servlet [dispatcherServlet] in context with path [] threw exception [Request processing failed: java.lang.NullPointerException: Cannot invoke "java.lang.Boolean.booleanValue()" because "java.util.List.get(int).active" is null] with root cause java.lang.NullPointerException: Cannot invoke "java.lang.Boolean.booleanValue()" because "java.util.List.get(int).active" is null at com.worldline.easypay.payment.control.PosValidator.isActive(PosVali dator.java:34) at com.worldline.easypay.payment.control.PaymentService.process(Paymen tService.java:46)
4. Observability
5. Alexandre TOURET Software Architect @touret.info blog.touret.info alexandre-touret David PEQUEGNOT Performance Engineer @foxlegend blog.worldline.tech foxlegend Who are we?
6. We design payments technology that powers the growth of millions of businesses around the world. 7000+ engineers in over 40 countries Managing 43+ billion transactions per year €250M spent in R&D every year Handling 150+ payment methods #1 European payment processor
7. Workshop
8. Pinpoint issues in the logs and improve them Broadcast metrics Implement Distributed tracing Correlating All of them Agenda
9. Pre-requisites github.com
10. https://worldline.github.io/observability-workshop https://github.com/worldline/observability-workshop The workshop
11. The EasyPay app
12. The EasyPay App
13. Under the hood poses t e s poses an c ec s O & card aut ori es t e payment customer
14. A fully observable platform
15. Logs, traces & metrics Logs Traces Metrics Where does it happen? Something happens What happens?
16. Think about observability from design Dashboards Alerts Logs Traces Metrics
17. poses an c ec s O & card aut ori es t e payment tores traces tores o s Data loading
18. OpenTelemetry Collector
19. n administrator o ro ses das ards and e tracts data from rafana tores traces tores o s Visualization n administrator o ro ses das ards and e tracts data from rafana tores traces tores o s
20. Instrumentation: Libraries or Agent ?
21. How to instrument my App? 2 ways Libraries • Faster at startup • Mandatory for native mode Agent • Works « out of the box » • Loose couple business artefacts from the technical ones • Smooth integration
22. Instrumentation Instrumentation or Agent Instrumentation OTEL Collector
23. From Libraries to agent Dashboards Alerts Logs Traces Metrics Java Agent
24. Observability & Performance
25. Logs Benchmark from Sebastian Daschner: Logging Performance Comparison https://www.javacodegeeks.com/2022/07/logging-performance-comparison.html Lack of logs impacts understanding of the application Too much logs impacts performances
26. Metrics Vert.X: Micrometer Metrics Performance https://vertx.io/blog/micrometer-metrics-performance/ Useful to understand the status of a platform The number of metrics can impact performance (e.g., Memory) •Be careful with histograms, buckets,... •Beware of the cardinality of the labels
27. Traces 1201,51 1477,82371 1468,98875 0 200 400 600 800 1000 1200 1400 1600 100% DISABLED 0% RPS (29 spans/trace) 1471,708845 1478,591626 1478,69221 0 200 400 600 800 1000 1200 1400 1600 100% DIS. 0% RPS (7 spans/trace) By default, a trace can be generated for each transaction But... this can have an impact on the performance of your application and storage → Use a collector
28. Traces Head Sampling & Performance 1477,82371 1468,98875 1422,40019 1320,57735 1252,411 1201,51 0 200 400 600 800 1000 1200 1400 1600 RPS - 29 spans/trace 549,89 553,13 571,25 615,37 648,82 676,22 0 100 200 300 400 500 600 700 800 Avg. Response Time (ms) – 29 spans / trace
29. Traces Head Sampling & Resources Usage 0,00% 10,00% 20,00% 30,00% 40,00% 50,00% 60,00% 70,00% 80,00% 90,00% 100,00% CPU% - 29 spans/trace 576,06 624,53 718,47 818,42 918,7 992,88 0 200 400 600 800 1000 1200 RSS (MB) - 29 spans/trace
30. Don’t e a stran er! Follow & get in touch @foxlegend linkedin.com/in /davidpequegnot Follow us: @touret.info linkedin.com/in /atouret 30 |
31. Explore our jobs in tech: careers.worldline.com Want to shape how the world pays and get paid?