Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Closing the door on application performance problems


Published on

Understand and analyze the reasons behind poor application performance. Learn how to measure end user satisfaction levels.

Published in: Technology
  • Be the first to comment

Closing the door on application performance problems

  1. 1. Closing the door on Application performance problems
  2. 2. 2. Analyzing transaction elements for performance issues 1. Reasons your application might be slow 3. Slow performing queries delaying response time 4. Measuring end user satisfaction levels Agenda
  3. 3. • A typical application infrastructure consists of different components that interact in a variety of ways to fulfill request from application end users • Web applications served from an application server • Database servers or data stores Application Model Overview
  4. 4. • Request Overload • Slice of death in Database • Application Code • Memory Management • Threads configuration 1. Reasons your application might be slow
  5. 5. Apdex score Application Performance Management Key performance indicators Application availability Error count Average Response time Request Rate Application & Server CPU
  6. 6. APM Insight for end to end visibility of your web applications
  7. 7. APM Insight Architecture
  8. 8. • Identify slow queries • Analyze exceptions • Comprehensive view on DB operations • Track custom methods • Code level diagnostics Deep Insights into your web application performance
  9. 9. Comprehensive Application performance using APM Insight
  10. 10. • Execute the installation package of APM Insight • Enter the IP Address of the applications manager instance • Configure other options such as apdex threshold, transaction tracing .Net Agent Installation
  11. 11. Customize agent configuration
  12. 12. • Global open standard • Measure of response time against a defined threshold • Scale of 0 to 1 • Satisfied, tolerated and frustrated users Measure end user satisfaction levels using apdex score
  13. 13. • Response time • Throughput • Slow business transactions • Error transactions 2. Analyzing transaction elements for performance issues
  14. 14. Individual business web transactions with component split up
  15. 15. • Individual business transaction apdex score • Transaction response time broken by components like web request, handle request, mvc, servlet, aspx • Transaction response time and throughput status Response time of individual business transaction elements
  16. 16. • Identify slow database calls • Find the most hit database tables, and determine if specific tables can be indexed well • Start tuning your database performance and diagnose slow queries 3. Slow performing queries delaying response time
  17. 17. See how your database is behaving overall
  18. 18. • Find the most executed database operations and its response time • Database operation by caller • Find the busiest table and the most performed SQL operation on the table. Individual database calls
  19. 19. • Identify slow internal invocations (methods) in the .NET code and trace the execution details for URLs Pinpoint slow application code
  20. 20. • Traces will chart the sequence of internal methods of the URL allowing you to drill down into transactions to view stack traces and SQL queries for slow or failed requests. Traces to investigate time consuming transactions
  21. 21. Trace Details contd
  22. 22. Drill down to SQL Statements
  23. 23. Details of background transactions
  24. 24. Pinpoint application errors
  25. 25. Stack trace of erroneous transactions
  26. 26. • Custom instrumentation allows you to monitor specific blocks of your application code. • It allows you to track your own application methods captured and displayed in the transaction traces. • The instrumented methods will be displayed with execution time under the trace details tab APM Insight Custom Instrumentation
  27. 27. Additional insights for Java web applications
  28. 28. • JVM Cpu usage • Heap & Non heap memory • Just in time compiler • JVM Classes count • Garbage collection details • Thread details • Configuration details JVM Performance insights
  29. 29. JVM Garbage collector
  30. 30. • Live, Daemon, Peak, Sleeping Threads • Aggregated CPU & user time • Deadlock threads JVM Metric threads
  31. 31. • The Health monitor is a self diagnostic tool, intended to provide a snapshot of the agent settings, for identifying and troubleshooting frequently encountered configuration issues in the agent APM Insight .Net agent health monitor • Check Network connectivity • Monitor worker processes • Check Application filters • View Service settings • Verify profiler status
  32. 32. • Capture errors, network requests, page load details and other metrics in your web applications • Immediately test on error transactions to eliminate false positives • Set up end user monitoring agents in branch facilities to check the availability and performance of your distributed applications 4. Measuring end user satisfaction levels
  33. 33. Understanding the end user journey
  34. 34. Installing EUM agents in your branch offices or customer locations
  35. 35. • Mimic user interaction • Detect and measure the critical components of each page with the help of waterfall model • Manage your business critical web flows like a pro • Reduce blame game between application development and network operations team Real Browser Monitoring
  36. 36. • Launch the web transaction recorder • Enter the host and SSL port of Applications Manager • Enter the Applications Manager credentials and proceed with the recording Recording a web script
  37. 37. Monitor critical business transaction workflows
  38. 38. Deeper visibility into multistep user interactions
  39. 39. • Use our EUM agent to understand global application performance through easy to use graphs and charts. Determine if performance is excellent, good, moderate, poor or bad when accessed from different countries. Measure web application performance across globe
  40. 40. Optimize the front end web components
  41. 41. • It depicts the series of actions that occur between a user and a server until a Web page is displayed in a end user’s browser. Find network latency using the waterfall model approach
  42. 42. • Provide seamless digital end user experience • Playback scripts including screen, keyboard, and mouse interactions performed by users • Meet SLAs by ensuring accurate performance of your web applications from all geographical locations • Pro active monitoring to reduce the number of reported incidents from end users • Playback scripts including screen, keyboard, and mouse interactions performed by users The need for end user monitoring
  43. 43. Contact : Online Demo : 1. Comprehensive capability to debug the enterprise grade business applications 2. Pinpoint application errors down to the line of code 3. Monitor JVM for Java applications 4. Measuring end user satisfaction levels of critical workflows across the globe Quick Recap:
  44. 44. Contact : Online Demo :