App Dynamics & SOASTA Testing & Monitoring Converge, March 2012


Published on

Dan Bartow, VP Performance Engineering, SOASTA and Steve Burton, Technology Evangelist, AppDynamics discuss the convergence of two traditionally separate domains and also demo testing and troubleshooting with CloudTest & AppDynamics.

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • Thank you so much for attending this webinar. The archive will be available on our website by the end of this week. Please tell others about it.
  • App Dynamics & SOASTA Testing & Monitoring Converge, March 2012

    1. 1. PERFORMANCE:Where Real-Time Testing & Production Monitoring Converge
    2. 2. In This Webinar:Reality Is Production…Let’s Test There.TODAY’S PRESENTERSSteve Burton: Technology Evangelist, AppDynamicsDan Bartow: VP Performance Engineering, SOASTABrad Johnson: VP Product Marketing, SOASTAAGENDAI. The convergence of two traditionally separate domainsII. Demonstration: Testing & Troubleshooting with CloudTest & AppDynamicsQuestions: Submit in the question box during event
    3. 3. Performance Issues Know No Borders Trouble spans teams, time and scale Exhausted Connection Pool Network Network configuration Capacity limits & Ops Hung JVM CDN file placement Conflict with other apps DNS routingStage / Team Load Balancer configuration OutOfMemoryException Auto-scaling failures Unbalanced web servers Max sockets exceeded Release & Deploy Slow third-party plug-ins s Synchronization Latency Security bottlenecks Caching issues and benefits Search technology limits Default configuration settings Garbage collection Inadequate server resources Reality? Any Memory leaks Memory leaks Database thread counts issue can occur Dev & Test Slow pages Inefficient database queries in production! Code bugs Method-level tuning Test Lab Staging Production (100% +++) Scale and Time
    4. 4. Old Way: Performance Testing Has Been a Lab Exercise DEPLOY Dev & Test Lab Production Environment Status Quo: Performance Testing Challenges:Small test window prior to release • Linear, multi-step approaches • Non-correlated, offline analysis • Disconnected teams Analyze • Costly test hardware • Tune Code • Build Test • Adjust • Non-scalable performance tools • Setup • Collect Environment Metrics Settings • Time-intensive setup • Find Issues Run Fix • Inability to hit realistic scale Time-consuming testing limits deployment options
    5. 5. Old Way: Problem Isolation Has Been a Lab Exercise DEPLOY Dev & Test Lab Production Environment Status Quo: Problem Isolation Challenges:Applications still crash and run slow in prod • Agile methodology increases change • SOA increases complexity • Data volumes differ from prod Analyze • Hardware specs differ from prod • Fix Code • Build Test • Optimize • Profilers slow application down • Setup • Response Environment Times Configuration • Reliability over Scalability • Exceptions Run Fix Make the application stable
    6. 6. What’s Changed?Agile development cycles compress test cycles• Days vs. weeks• Incomplete knowledge transfer from Dev to Test to Ops• Difficult to simulate every user session variation Production environments are difficult to replicate • Horizontal scaling requires many commodity servers • Transaction volumes have Increased with Web & mobile commerce • Large data sets are difficult to replicate Costly to simulate peak loads in test environments • Performance testing & optimization need to change, too…
    7. 7. New Way: Test & Monitor your Applications from Lab into Production Dev & Test Lab Production EnvironmentStart small, Isolate & fix Hit expected Detect Recheck Solve issuesearly to root performance scale in problems in regularly for as they occur out code- issues in the staging and both real and regression in thelevel issues lab maintenance synthetic and capacity lifecycle windows sessions
    8. 8. Everyone Wants to be Agile Old Way New Way (dev/test) (production)Load Testing Weeks HoursIsolate Problems Hours Minutes Traditional Approaches TEST BUILDING HW Setup TEST ANALYZE FIX RETEST REPEAT? TIME BUILD H W TEST H W TEST H W TEST H W TEST H W TEST Extra ANALYZE ANALYZE ANALYZE ANALYZE ANALYZE S FIX S S S S time! FIX FIX FIX FIX e e e e e t t t t t u p u p Agile Performance Approach u p u p u p
    9. 9. “But, Testing in Production Is SCARY!”(Isn’t it?) It is when Key Criteria eliminate never intended for These using tools & approaches RISK & make Production Testing a critical requirement! the complexity & magnitude of modern environments CONTROL – Stop, Start, Modify, Pause user traffic as tests run REAL TIME ANALYSIS – access to all performance info as issues occur SCALE – simulate any number of geographically disperse usersSPEED – able to deploy and gaininsight within minutes
    10. 10. The Convergence of Real-Time Testing & Production Monitoring is HERE NOW!Demonstration – SOASTA & AppDynamics Dan Bartow Steve Burton (SOASTA) (AppDynamics)
    11. 11. Q&A Try us for FREE: Questions & Answers will post on CloudLink: Previous & Upcoming Webinars: Contact SOASTA: Contact AppDynamics 866.344.8766 415-442-8400 Follow us: Follow us: