Doing DevOps? Better base it on performance - DevNexus 2014

971 views

Published on

DevOps is a proven practice for high performing IT organizations. It is the new buzzword in IT but larger enterprises are having mixed if any success. Culture change can be a daunting task especially since most of the focus is around operations. Today its not just about the end user’s experience but also about how quickly the business can deliver new and innovative features to the market. This is hard when the two sides of this delivery model are opposed to each other. Operations wants stability and development wants velocity. With this type of constraint on time and resources the idea of doing right is shadowed by doing it fast. In this talk I would like to share how development and operations can come together using performance as the currency to drive innovation. Performance is the one constant that everyone can use to apply DevOps practices with high adoption results.

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
971
On SlideShare
0
From Embeds
0
Number of Embeds
21
Actions
Shares
0
Downloads
10
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Doing DevOps? Better base it on performance - DevNexus 2014

  1. 1. Stephen  Wilson   • apmblog.compuware.com   • @steve0wilson  (0  as  in  zero)   • LinkedIn   • stephen.wilson@compuware.com   1   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  2. 2. 80% of a Development Budget is spent on ID and fixing defects 60 150 2   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  3. 3. 3   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  4. 4. 80% of a Development Budget is spent on ID and fixing defects 60B dollars is the cost of defects that reach production 150 4   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  5. 5. 5   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  6. 6. 80% of a Development Budget is spent on ID and fixing defects 60B dollars is the cost of defects that reach production 150 X the cost to fix a problem in production vs. in development 6   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  7. 7. 7   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  8. 8. DOING  DEVOPS?     WHY  YOU  NEED  TO   FOCUS  ON   PERFORMANCE   Andreas  Grabner,  Technology  Strategist  @  Compuware/dynaTrace   8   8  
  9. 9. 9   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  10. 10. 10   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  11. 11. 11 11  
  12. 12. 12   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  13. 13. 13 13  
  14. 14. The Waterfall Process Development   TesZng   ProducZon  Release   14   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  15. 15. Then came Agile… 15 15  
  16. 16. Rapid Acceleration… 16   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  17. 17. Production Problems 17   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  18. 18. Produc>on   (local,  remote,  outsourced)   •  Cloud  load  tesZng   •  Monitoring   18 18  
  19. 19. 19   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  20. 20. process? buzzword? a way of life? agile ops? 20   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  21. 21. Seeking  Knowledge  vs.  Applying  Knowledge   21   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  22. 22. BUSINESS  ALIGNMENT   LEGACY  ATTITUDES CHALLENGES   TRADITIONS   NATURAL  FRICTION   ORGANIZATIONAL  TRUST   22   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  23. 23. ReacZve   23   ProacZve   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE PredicZve  
  24. 24. 24   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  25. 25. 25   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  26. 26. 26   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  27. 27. 27   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  28. 28. •  Jprobe   •  Eclipse   •  Debuggers     Development   (local,  remote,  outsourced)   28   •  Load  TesZng   •  App  Monitoring   Test/QA   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE •  •  •  •  Ops  tools   Network  Tools   Log  scraping   App  Monitoring   Produc>on  
  29. 29. Problem:  Unclear  Ops  Problem  Descrip>ons   29   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  30. 30. 30   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  31. 31. Unclear  End  User  Problem  Descrip>ons  …   31   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  32. 32. …  when  you  really  want  to  have  this  …   32   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  33. 33. …  and  this   Top  Contributor  is  related  to   String  handling   99%  of  that  Zme  comes  from   RegEx  Pabern  Matching   Page  Rendering  is  the  main  component   33   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  34. 34. 34   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  35. 35. STOP  THE  BLEEDING   COMMUNICATE  THE  WINS   IDENTIFY  KEY  PLAYERS   35   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  36. 36. 36   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  37. 37. 37   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  38. 38. 38   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  39. 39. 39   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  40. 40. 40   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  41. 41. Performance  Focus  in  Test  Automa>on   Embed  your  Architectural  Results   in  Jenkins   41   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  42. 42. Performance  Focus  in  Test  Automa>on   Build-­‐level   Performance  Metrics     42   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  43. 43. STANDARDIZE  THE  PROCESS   DEFINE  PERFORMANCE  DRIVERS   CREATE  A  CONSTITUTION   43   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  44. 44. 44   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  45. 45. How?  Performance  Focus  in  Test  AutomaZon   Lets look behind the scenes Architectural  Data   Test  Framework  Results   Build  #   Test  Case   Status   Build  17   testPurchase   OK   12   0   120ms   testSearch   OK   3   1   68ms   testPurchase   FAILED   12   5   60ms   testSearch   OK   3   1   68ms   testPurchase   OK   75   0   230ms   testSearch   OK   3   1   68ms   Build  18   Build  19   Build  20   45 45   #  SQL   #  Excep   CPU   12   0   120ms   Exceptions probably reason 3   1   68ms   testSearch   identified a regresesion OK   We for failed tests Problem fixed but now we have an Problem solved architectural regression Now we have the functional and architectural confidence testPurchase   OK  
  46. 46. How?  Performance  Focus  in  Test  AutomaZon   Analyzing  All  Unit  /   Performance  Tests   IdenZfy   Regressions   Analyze  Perf   Metrics   46 46  
  47. 47. How?  Performance  Focus  in  Test  AutomaZon   Cross  Impact  of  KPIs   47 47  
  48. 48. 48   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  49. 49. AUTOMATE  THE  PROCESS   DRIVE  FEEDBACK   PERFORMANCE  BASED  DECISIONS   49   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  50. 50. ROME  WAS  NOT  BUILT  IN  A  DAY   SPONSORSHIP   DON’T  BE  AFRAID  TO  FAIL   TALENT   KNOWLEDGE   50   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  51. 51. Stephen  Wilson   • apmblog.compuware.com   • @steve0wilson  (0  as  in  zero)   • LinkedIn   • stephen.wilson@compuware.com   51   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE
  52. 52. Thank  You   ParZcipate  in  Compuware   APM    Discussion  Forums   apmcommunity.compuware.com   Like  us  on  Facebook   Follow  us  on  TwiTer   twiber.com/CompuwareAPM   Read  our  Blog   facebook.com/CompuwareAPM   About:Performance   Join  our  LinkedIn  group   Watch  our  Videos  &     product  Demos   Compuware  APM  User  Group   youtube.com/Compuware   www.compuware.com/APM   52 52   © 2011 Compuware Corporation — All Rights Reserved
  53. 53. 53   COMPANY CONFIDENTIAL – DO NOT DISTRIBUTE © 2011 Compuware Corporation — All Rights Reserved

×