Elastic{ON} 2018 - Visualizing Content Performance with Elastic at Canadian Broadcasting Corporation (CBC)

Transcript

1. Canadian Broadcasting Corporation February 28th, 2018 Visualizing Content Performance with

   Elastic Sean Bennett & Stefan Piruzevski

2. In brief We are going to talk about… - Some

   history - What brought us here - Our challenge

3. Who we are Stefan Piruzevski Sean Bennett

4. A little history

5. Our mandate • Distinctly Canadian • Contribute to cultural expression

   • Shared national consciousness and identity • Meet Canadians where they consume content • Reflect the cultural of Canada

6. We’re local.

7. Like, really local ENGINEERING PHOTO FROM RESOLUTE BAY

8. Which is big!

9. Getting our house in order

10. What we deliver Vertical focused content (News, Comedy, Lifestyle, Sports)

11. How do we deliver it?

12. We’re a Broadcaster

13. The Ask - Where it all started (SP) Number of

    stories updated per day, per location, per particular time of day

14. Educated guess with rudimentary numbers

15. mvn exec:java -Dexec.mainClass="ca.cbc.INTERNAL.PATH.util.ReferringContent" -Dexec.args="http://<internal path>:<port>/connection-properties/conn ection.properties sysadmin sysadmin output.csv 'inputTemplate:Image

    AND (text:Flickr OR text:Flicker)'" Done. Documents found: 421 Success: 421 Failed: 0 Referring content ids: 533 Sample Report!

16. CBC’s Infrastructure

17. Content Management System

18. Owned and Operated

19. Owned and Operated

20. Owned and Operated

21. Owned and Operated

22. System map

23. System map CDN (Cache) Internal (Cache) CMS (Cache) CMS (Disk)

    CMS (DB)

24. We looked at:

25. We looked at: building a custom solution

26. We looked at: logging tools used by our network teams

27. We looked at: other vendors and CMS platforms

28. We looked at: existing internal tools

29. We looked at: creating more JAVA classes

30. We looked at: The possibility of creating a mirrored environment

    with real-time replication of our data so we could query the db in a non-audience facing environment.

31. We looked at: EVERYTHING

32. You said “Banana”!

33. BANANAS! Query existing Solr indexes Graphical UI!!! No more manual

    Solr queries We loved it Performance vs Flexibility
34. None

35. Dynamic provisioning Not complicated Easily incorporate into our CMS

36. None

37. What are we using

38. What are we using To Logstash or not to Logstash

    Real-time data with Beats Easily add Filebeat into our CMS

39. What are we using Easy to learn Already used in

    by some teams Not directly connected to Production

40. What are we using Bad Data Wiping Everything Duplicate Data

41. What are we using

42. What are we using Apply Banana Lessons Real-time Dashboards Don’t

    judge us.

43. What are we using

44. Let’s Assemble! ?

45. The players !

46. None

47. The Ask - Where it all started (SP) Number of

    stories updated per day, per location, per particular time of day

48. Number of stories updated per day, per location, per particular

    time of day

49. Now what? ?

50. Capture all the things!

51. Number of stories updated per day, per location, per particular

    time of day

52. Number of stories updated per day, per location, per particular

    time of day

53. Number of stories updated per day, per location, per particular

    time of day

54. Number of stories updated per day, per location, per particular

    time of day
55. None
56. None
57. None
58. None
59. None
60. None
61. None
62. None
63. None
64. None

65. Rinse, Repeat

66. Next level data-driven decision making •Workflow/Process improvements •Identifying opportunities •Giving

    more accurate performance measurement

67. Let’s loop in the audience!

68. Thanks! Catch us at the AMA booth! Stefan Piruzevski Sean

    Bennett