CI/CD with Kubernetes

Transcript

1. @BastianHofmann
   CI/CD with Kubernetes
   Bastian Hofmann
   mail@bastianhofmann.de
   

   View Slide

2. Continuous Integration
   

   View Slide

3. Continuous Delivery
   

   View Slide

4. View Slide

5. Container orchestration platform
   

   View Slide

6. Deploy, run and scale your services
   in isolated containers
   

   View Slide

7. No vendor lock in
   

   View Slide

8. Standardized APIs
   

   View Slide

9. Runs on
   

   View Slide

10. Your laptop
    

    View Slide

11. View Slide

12. Bare metal
    

    View Slide

13. Cloud Providers
    

    View Slide

14. And if you don't want to install and
    maintain Kubernetes yourself
    

    View Slide

15. Managed Kubernetes
    

    View Slide

16. View Slide

17. View Slide

18. So we have code, how do we get this
    into Kubernetes?
    

    View Slide

19. Lot's of tools
    

    View Slide

20. Example for this talk
    

    View Slide

21. View Slide

22. Example application
    

    View Slide

23. https:/
    /gitlab.com/bashofmann/
    angular-test-app
    

    View Slide

24. https:/
    /gitlab.com/bashofmann/
    angular-test-app-finished
    

    View Slide

25. Pipeline with multiple stages, each
    stage can have multiple jobs
    

    View Slide

26. Stage 1:
    Preparing
    

    View Slide

27. Stage 2:
    Linters and Tests
    

    View Slide

28. Stage 3:
    Building an image
    

    View Slide

29. Stage 4: On master:
    Deploying to Kubernetes into a stage
    namespace
    

    View Slide

30. Stage 5: On master:
    Testing if stage works
    

    View Slide

31. Stage 6: On tag:
    Tag the docker image
    

    View Slide

32. Stage 7: On tag:
    Deploying to Kubernetes into a prod
    namespace
    

    View Slide

33. Stage 8: On tag:
    Testing if prod works
    

    View Slide

34. In detail
    

    View Slide

35. Stage 1:
    Preparing
    

    View Slide

36. Only do slow operations once and
    cache and re-use results
    

    View Slide

37. Stage 2:
    Linters and Tests
    

    View Slide

38. Add linters for everything
    

    View Slide

39. Unit tests
    

    View Slide

40. Integration tests
    

    View Slide

41. Browser tests
    

    View Slide

42. Stage 3:
    Building an image
    

    View Slide

43. Make the build as fast as possible
    

    View Slide

44. Make the image as small as possible
    

    View Slide

45. Stage 4: On master:
    Deploying to Kubernetes into a stage
    namespace
    

    View Slide

46. Stage 5: On master:
    Testing if stage works
    

    View Slide

47. Stage 6: On tag:
    Tag the docker image
    

    View Slide

48. Stage 7: On tag:
    Deploying to Kubernetes into a prod
    namespace
    

    View Slide

49. Stage 8: On tag:
    Testing if prod works
    

    View Slide

50. Use a versioning scheme
    

    View Slide

51. Never deploy "latest"
    

    View Slide

52. Use Kubernetes namespaces
    

    View Slide

53. Use Helm or similar tools
    

    View Slide

54. Make use of Kubernetes "magic"
    

    View Slide

55. Sealed Secrets
    

    View Slide

56. Secrets are stored encrypted in Git
    and encrypted inside of the cluster
    

    View Slide

57. LoadBalancers
    

    View Slide

58. Automatically creates a highly
    available LoadBalancer with a
    public IP address
    

    View Slide

59. Ingress controller
    

    View Slide

60. The ingress controller (nginx) listens
    on routing rules in Ingress Resources
    and configures itself to route
    incoming traffic to the correct
    running and healthy pods
    

    View Slide

61. cert-manager
    

    View Slide

62. Cert-manager listens on Ingress
    Resources and if they need TLS,
    requests a certificate from
    LetsEncrypt
    

    View Slide

63. external-dns
    

    View Slide

64. External-DNS listens on Ingress
    Resources and creates DNS entries
    

    View Slide

65. Prometheus
    

    View Slide

66. Kubernetes and Node metrics are
    automatically included
    

    View Slide

67. Kubernetes and Node alerts are
    automatically included
    

    View Slide

68. Kubernetes and Node dashboards
    are automatically included
    

    View Slide

69. Scraping targets can be
    automatically discovered
    

    View Slide

70. Service Meshes
    

    View Slide

71. View Slide

72. Alternatives
    

    View Slide

73. Flux
    

    View Slide

74. View Slide

75. Knative
    

    View Slide

76. Kubernetes-based platform to
    deploy and manage modern
    serverless workloads
    

    View Slide

77. Tekton CD
    

    View Slide

78. CI that runs completely within
    Kubernetes with Kubernetes Custom
    Resources
    

    View Slide

79. mail@bastianhofmann.de
    https:/
    /twitter.com/BastianHofmann
    http:/
    /speakerdeck.com/u/bastianhofmann
    

    View Slide