Faster apps. faster time to market. faster mean time to repair

630 views

Published on

Developers, Test Engineers, QA Engineers, Network Engineers, Operations Managers, Production Managers and Solution Architects joined us in Singapore to learn more about APM Lifecycle

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
630
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
3
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Faster apps. faster time to market. faster mean time to repair

  1. 1. Faster Apps, Faster Time to Market,Faster Mean Time to RepairBrad GoddardDirector of APM Pre-Sales Engineering - Asia and IndiaCompuwareArdeshir ArfaianSolution Director dynaTrace APACCompuware
  2. 2. Compuware Application Performance Management We help organizations optimize the performance of their business-critical applications • Web, non-Web, mobile, streaming, cloud-based applications • Across all customers, users, browsers, devices, infrastructure, and geographies • Rapid issue notification with actionable diagnostics • Insight into how these issues affect your business (revenue, brand, cost) SaaS, 4,000+ Customers Global Reach Recognized as Cloud-Based and Worldwide • Over 80 offices in Industry Leader On-Premises • 2,500+ enterprise 29 countries • Gartner: Offerings customers worldwide Leader in APM magic • Rapid startup and • 1,500+ SMB • Strategic service quadrant payback customers delivery • Forrester Research: • 12 of top 20 “…a complete view of US sites end user experience”* • Ovum: “Game-changing”*”Trends: The Diversification Of End User Experiencing Monitoring”, Forrester Research, Inc., July 5, 2011
  3. 3. Your world is changing Application visibility and optimization of the customer experience are more important than ever.Customers: Global New Devices: ProliferatingApplications: Distributed and loosely coupled Virtualization/Cloud: Exploding
  4. 4. Impact of to the business
  5. 5. The Problem Lifecycle
  6. 6. Why Agile Development took off
  7. 7. Story Points It‘s Sprint Time! Development Testing Estimate Sprint Timeline Remaining Production Team Velocity
  8. 8. Story Points You are in control! Developme Testing Estimate nt Sprint Timeline Remaining Production Team Velocity
  9. 9. Story Points What happened? Developme Testing Estimate nt Sprint Timeline Remaining Production Team Velocity
  10. 10. Story Points Missed Goals and Estimates Missed Developme Testing Estimate nt Production Missed Remaining Goal Team Velocity
  11. 11. 4 of 5 projects run over time and/or budget. Oxford University Regarding ITÂ Project Success (Saur & Cuthbertson, 2003) 11
  12. 12. Problem #1: Different MindsetSource: http://dev2ops.org/blog/2010/2/22/what-is-devops.html
  13. 13. Problem #2: Dislocated TeamsSource: http://dev2ops.org/blog/2010/2/22/what-is-devops.html
  14. 14. Problem #3: Different ToolsSource: http://dev2ops.org/blog/2010/2/22/what-is-devops.html
  15. 15. Problem #4: Over the Fence AttitudeSource: http://dev2ops.org/blog/2010/2/22/what-is-devops.html
  16. 16. These Problems lead to …Source: http://dev2ops.org/blog/2010/2/22/what-is-devops.html
  17. 17. A potential SolutionSource: http://dev2ops.org/blog/2010/2/22/what-is-devops.html
  18. 18. Real World Perf Test in Feedback CI ONECloud based Toolset Architecture ValidationTesting Test in Production Traditional Load Testing
  19. 19. Minimize and automate real Load TestsDeveloping Test Run Reproduction Refine Capturing Re Run Tests Reproduction Refine Capturing Multiple Test Iterations needed to analyze Root-cause Re Run Tests Reproduction Problem Analysis Problem Solving timeDeveloping Test Run Reproduction Refine Capturing Re Run Tests Reproduction Refine Capturing •Eliminates Test Iterations •Go directly to problem analysis •Frees up resources for other proje Re Run Tests Reproduction Problem Analysis Problem Solving time
  20. 20. Why Web Performance Matters: Impact of PoorPerformance found that a 2 second slowdown 4.3 % reduction in revenue/user* determined that a 400 millisecond delay 0.59 % fewer searches/user* Source: Steve Souders @ Velocity Conference 2009 http://radar.oreilly.com/2009/07/velocity-making-your-site-fast.html
  21. 21. 21
  22. 22. ….10000 Smart Phones Sold 22
  23. 23. ….80000 electronicaccessories sold 23
  24. 24. eBay Marketplace = Economy of Scale 22B 10B 10B page views/day URL Requests / day40 40M 9 9 Petabytes of data storage $62 $62B lines of code 2010 gross merchandise volume100 5 300100M 300M active users live listings10000 75 10,000 5K 75B search engine nodes application servers database calls/day 24Commercial data warehouse 100x larger than the research library ofUS Congress
  25. 25. Pertinent Problems to be solved @ eBay• Search• Trust, Fraud and Risk• Shipping and Logistics• Ease of Payments• User Experiences & Site Speed• Data , Analytics and Business Intelligence• Performance …• … and many more 25
  26. 26. Benchmark CriteriaS No eBay Requirements Status 1 Deeper insight into the application very quickly, identifying the areas of  code where the majority of each transactions time is spent. 2 Integrate with Silk Performer / JMeter  3 Java Diagnosis at method/class level.  4 API Breakdown chart  5 Memory Analysis graph  6 Dashboard showing a comparison between 2 different test runs  7 Trace export for QA, Dev  8 Business and Technical dashboards  9 Execution time / Time spent in individual methods of the Application  code base 10 Time Spent on Service calls. (Entry/Exit times only)  11 Performance of SQL Queries.  26 12 Reports that would help identify the slow parts of the Application  13 To be able to configure and monitor performance of specific business  flows.
  27. 27. Link to Compuware APM 27
  28. 28. Selected transactions opensin Compuware APM 28
  29. 29. How much time is spent on which tier? 29Are all my tiers healthy?
  30. 30. Detailed view of transaction and flow Each individual transaction listed Selected transaction spent 42.77 milliseconds Layers Transaction spent time in 30
  31. 31. API level Drill down toidentify the method and the call path havingmaximum performance impact 31
  32. 32. Global Solution ProviderFinancial Services
  33. 33. Transaction Breakdown<1sec, 1-2sec, 2-3sec, 3-4sec, 4-5sec, >5sec With increasing load number of Outliers >5sec is increasing
  34. 34. Only 85.44% of transactions under 1 secondGoal is to have 90% of transactions under 1 second.
  35. 35. High Connection Checkin/Checkout time High RMI execution time
  36. 36. JDBC Connection Check-in/Check-out (1) High Avg wait time for a connection (10 seconds)
  37. 37. Low CPU / Low Memory consumption / High GC Memory Utilization never climbs above 25 % on certain JVMs. Even though GC is high.
  38. 38. High GC JVM is spending 5.75 minutes per minute on GC
  39. 39. GC versus Exec Time ratiocommon.dbservices JVM is spending 96% of it’s time on GC Further analysis showed that most of GC time are major GCs
  40. 40. Root-Cause JVM is running in Client mode
  41. 41. GC versus Exec Time ratiocommon.dbservicesAfter switching JVM to servermode, GC time is drastically reduced.Further analysis showed only minor GCs
  42. 42. Before (client-mode JVM) / After (server-mode JVM) SLA levels restored With increasing load number of Outliers >5sec is increasing Moving production load to other datacenter & applying –server option in meantime
  43. 43. Innovation……and Getting Acquired
  44. 44. Faster Apps, Faster Time to Market,Faster Mean Time to RepairBrad GoddardDirector of APM Pre-Sales Engineering - Asia and IndiaCompuwareArdeshir ArfaianSolution Director dynaTrace APACCompuware

×