Path to SRE

You’ve heard all the buzz about SRE, but what does it actually take to do it if you are not Google? In this talk you will learn about our experience creating Auth0's SRE flavor and rolling it out.

Transcript

1. The Path to SRE
   @dschenkelman
   Director of Engineering @auth0
   

   View Slide

2. SRE
   

   View Slide

3. Why?
   

   View Slide

4. Reliability is the one feature
   every customer uses
   - an @auth0 SRE
   

   View Slide

5. Auth0
   User
   Auth0
   Customer App
   

   View Slide

6. Context
   

   View Slide

7. Focused Investment
   Like Security but for Reliability
   

   View Slide

8. Scale
   

   View Slide

9. Research
   

   View Slide

10. Companies
    

    View Slide

11. Organizations
    

    View Slide

12. Style
    

    View Slide

13. Sponsors
    

    View Slide

14. Who?
    

    View Slide

15. Spectrum
    Systems Software
    

    View Slide

16. The Usual Suspects
    

    View Slide

17. Teachers
    

    View Slide

18. Advocates
    

    View Slide

19. Problem solvers
    

    View Slide

20. Know the system
    

    View Slide

21. Experience
    

    View Slide

22. node.js
    

    View Slide

23. Educate
    

    View Slide

24. What we do
    SRE identifies, develops, refines,
    and disseminates the libraries,
    services, practices, and processes
    key to system reliability.
    

    View Slide

25. SRE does not force itself
    on other teams
    

    View Slide

26. SRE does not handle all
    incident response
    

    View Slide

27. Involvement Spectrum
    SRE Run Service
    Embedding
    Consultancy
    Office Hours/
    Workshops
    

    View Slide

28. Contacting SRE
    

    View Slide

29. The brand
    

    View Slide

30. Logo
    

    View Slide

31. Office Hours
    

    View Slide

32. Brown bags
    

    View Slide

33. Investigations
    

    View Slide

34. Flexibility
    

    View Slide

35. Incidents
    

    View Slide

36. Execute!
    

    View Slide

37. You are selling TRUST
    

    View Slide

38. SLOs
    

    View Slide

39. R2
    

    View Slide

40. View Slide

41. View Slide

42. View Slide

43. Incident Response
    

    View Slide

44. Distributed Traces
    

    View Slide

45. Rate limiting
    

    View Slide

46. CI/CD
    

    View Slide

47. Complex Issues
    

    View Slide

48. Today
    

    View Slide

49. Org
    IAM DX
    Platform
    SRE
    

    View Slide

50. Results
    • 5/11 teams doing R2s organically
    • > 5x more frequent deploys with < 10x
    duration
    • 80% critical services with tracing
    

    View Slide

51. Results (2)
    • 5 complex issues solved
    • > 99.99% reliability for User
    Management API
    • ~8ms -> ~3ms 99th perc latency for
    rate limits
    

    View Slide

52. Success
    

    View Slide

53. Vision
    Subject to change :)
    IAM DX
    Platform
    SRE
    PR
    SRE
    AR
    SRE
    AR
    SRE
    OX
    

    View Slide

54. Thanks
    @dschenkelman
    

    View Slide