Q Con Performance Testing At The Edge
1 like483 views
This document discusses performance testing at the edge using dynaTrace. It describes challenges with traditional approaches like waterfalls and profiling, which don't provide ongoing monitoring and comparison over time. The dynaTrace approach uses a distributed architecture to monitor many platforms and configurations. It also discusses lessons learned around making testing continuous, avoiding re-runs, and focusing tests on specific metrics like response time and garbage collection to understand measurements better.
More Related Content
Similar to Q Con Performance Testing At The Edge (20)
More from Alois Reitbauer (20)
![Strategic Picks — Prioritising the Right Agentic Use Cases [2/6]](https://cdn.slidesharecdn.com/ss_thumbnails/session-2-250902135931-4c97717a-thumbnail.jpg?width=560&fit=bounds)
Q Con Performance Testing At The Edge
- 1. Performance Testingatthe EdgeAlois Reitbauer, dynaTrace Software
- 7. ManyplatformsDifferent usagescenariosHigh numberofconfigurationsNo easy waytopatchsoftware
- 8. 8OurArchitectureDYNATRACE CLIENT DYNATRACE SERVERDYNATRACE COLLECTOR(OPTIONAL)DYNATRACE COLLECTOR(OPTIONAL)WANJava Server.NET ServerDatabase Web ServerAPPLICATION
- 10. Profiling was not enoughGoodforfindingproblemsResultcomparisonhardOnly valid untilnextcheck-inToomuchwork
- 11. The Life of a Log StatementEnterthecodepublicvoidfoo (){ … // do something bar ();}publicvoid bar (){ … // do somethingelse Logger.log (“executiontook“ + time);}
- 12. The Life of a Log StatementSomebodychangessomethingpublicvoidfoo (){ … // do somethingif (condition) bar ();elseotherBar ();}publicvoid bar (){ … // do somethingelse Logger.log (“executiontook“ + time);}
- 13. The Life of a Log StatementYourcodegetsdeprecatedpublicvoidfoo (){ … // do somethingnewBar ();}publicvoid bar (){ … // do somethingelse Logger.log (“executiontook“ + time);}
- 14. Methodology
- 15. DefiningourstrategyStartearlyBreak in piecesTest Continuously
- 16. Frequency vs.GranularityJUnit-basedTests (2x day)Total SystemTestsLong-runningStabiltiy Tests(2 w duration)GranularityFrequency
- 18. Avoid Re-RunsWhatcould happen?Whichinformation do youwant?Whatdescribesyoursystem?Whatis different fromthe last run?
- 19. Aim high …… test50%more
- 22. Measurements
- 25. Trends in Unstable Tests
- 27. UnderstandyourmeasurementsResponse Time and GCResponse Time only
- 28. BeSpecific on whattotestThroughputResponse TimeMemory ConsumptionOther KPI …
- 29. Beyond Response TimeKPI Chart: Server Throughput Over Time
- 30. Motivateyourteam
- 31. Howtomakedeveloperswritetests#1 Heroism#2 Boomerang#3 The otherguy#4 Bug me not#5 Feedback#6 Code vs. Wine#7 Newb vs. Noob
- 32. Test Case ComplexityFirstStart dynaTrace infrastructureWhenreadyStart n WebSphereinstances on servers … WhenreadyStart LoadtestagainstWebSphereserversAfter loadteststartExecute testcase
- 33. Making complexthings easy   @DtdRemoteSud(           host = "lab2",           name = "WAS7.0",           startupPriority = 1,           postStartClosure = WaitForWebSphereSudIsUp.class   )   private SudInterfacewebSphereSud;
- 37. [email protected] Mailblog.dynatrace.com BlogAloisReitbauerTwitter