What Even is Cloud Native? (PHPNE, April 2018)

What Even is Cloud Native? (ScotlandPHP 2017) by David McKay

Published November 7, 2017 in Technology

The Cloud Native Computing Foundation, the organisation currently governing Kubernetes, Prometheus, OpenTracing, and more; describe the three goals of "Cloud Native" applications as:

Containerized
Each part (applications, processes, etc) is packaged in its own container. This facilitates reproducibility, transparency, and resource isolation.

Dynamically orchestrated
Containers are actively scheduled and managed to optimize resource utilization.

Microservices oriented
Applications are segmented into microservices. This significantly increases the overall agility and maintainability of applications.
In this talk, I will guide you towards taking your application cloud native, utilising the software available to us today, from the CNCF, and others, covering containers, tracing, logging and service discovery ... as well as the dreaded: "What actually is a micro-service?"

Transcript

1. @rawkode
   What Even is Cloud Native?
   

   View Slide

2. @rawkode
   David McKay
   Software
   Consultant
   ➔ User Group Organiser
   ◆ Cloud Native Glasgow
   ◆ Docker Glasgow
   ◆ DevOps Glasgow
   ◆ Pair Programming Glasgow
   ◆ MongoDB Glasgow
   ➔ Developer
   ◆ Ballerina
   ◆ Crystal
   ◆ Elixir
   ◆ Go
   ◆ PHP
   ◆ Python
   ◆ Pony
   

   View Slide

3. @rawkode
   Cloud Native
   Klaʊd Neɪtɪv
   adjective: cloud-native, superlative adjective: cloud-nativest
   ???
   

   View Slide

4. @rawkode
   Cloud
   Klaʊd
   

   View Slide

5. @rawkode
   Cloud
   Klaʊd
   

   View Slide

6. @rawkode
   Cloud
   Klaʊd
   * To all the OCD types out there, I’m sorry
   

   View Slide

7. @rawkode
   Cloud Native Computing Foundation
   (CNCF)
   

   View Slide

8. @rawkode
   Cloud Native Computing Foundation (CNCF)
   1. Containerized
   Each part (applications, processes, etc) is packaged in its own container. This facilitates
   reproducibility, transparency, and resource isolation
   2. Dynamically Orchestrated
   Containers are actively scheduled and managed to optimize resource utilization
   3. Microservices Oriented
   Applications are segmented into microservices. This significantly increases the overall agility
   and maintainability of applications.
   

   View Slide

9. @rawkode
   Containerized
   “Each part (applications, processes, etc) is packaged in its own container. This
   facilitates reproducibility, transparency, and resource isolation”
   

   View Slide

10. @rawkode
    Containerized: Reproducibility
    

    View Slide

11. @rawkode
    Containerized: Resource Isolation
    

    View Slide

12. @rawkode
    Containerized: Transparency
    

    View Slide

13. @rawkode
    Containerized: Transparency
    FROM alpine:3.5
    RUN apk update && apk add php
    ENTRYPOINT [ “php” ]
    CMD [“-v”]
    

    View Slide

14. @rawkode
    Dynamically Orchestrated
    Containers are actively scheduled and managed to optimize resource utilization
    

    View Slide

15. @rawkode
    Dynamically Orchestrated
    

    View Slide

16. @rawkode
    --memory=4M
    --memory-reservation=4M
    --cpus=1.5
    --cpu-period=100000 --cpu-quota=50000
    

    View Slide

17. @rawkode
    spec.containers[].resources.limits.cpu
    spec.containers[].resources.limits.memory
    spec.containers[].resources.requests.cpu
    spec.containers[].resources.requests.memory
    

    View Slide

18. @rawkode
    Microservices Oriented
    Applications are segmented into microservices. This significantly increases the
    overall agility and maintainability of applications.
    

    View Slide

19. @rawkode
    // npm: is-even:1.0.0
    if (i % 2 == 0) {
    return true;
    }
    return false;
    

    View Slide

20. @rawkode
    // npm: is-even:1.0.0
    var isOdd = require('is-odd');
    module.exports = function isEven(i) {
    return !isOdd(i);
    };
    

    View Slide

21. @rawkode
    // is-odd:2.0.0
    var isNumber = require('is-number');
    module.exports = function isOdd(i) {
    if (!isNumber(i)) {
    throw new TypeError('is-odd expects a number.');
    }
    if (Number(i) !== Math.floor(i)) {
    throw new RangeError('is-odd expects an integer.');
    }
    return !!(~~i & 1);
    };
    

    View Slide

22. @rawkode
    Microservices Oriented
    

    View Slide

23. @rawkode
    I’m sold. What do I need?
    

    View Slide

24. @rawkode
    Cloud Native Computing Foundation (CNCF)
    ★ Kubernetes
    ★ Prometheus
    ★ OpenTracing
    ★ Fluentd
    ★ Linkerd
    ★ gRPC
    ★ CoreDNS
    ★ containerd
    ★ rkt
    ★ CNI
    ★ Envoy
    ★ Jaeger
    

    View Slide

25. @rawkode
    Kubernetes
    Orchestration
    

    View Slide

26. @rawkode
    Kubernetes
    Orchestration
    ➔ ConfigMaps
    ➔ Secrets
    ➔ Pods
    ➔ Deployments
    ➔ StatefulSets
    

    View Slide

27. @rawkode
    apiVersion: v1
    kind: ConfigMap
    metadata:
    name: my-configmap
    data:
    api_domain: "example.com"
    

    View Slide

28. @rawkode
    apiVersion: extensions/v1beta1
    kind: Deployment
    spec:
    replicas: 3
    template:
    spec:
    containers:
    - name: my-container
    image: rawkode/my-container:sha256
    

    View Slide

29. @rawkode
    kind: Service
    apiVersion: v1
    spec:
    selector:
    name: my-deployment
    type: NodePort
    ports:
    - protocol: TCP
    port: 80
    

    View Slide

30. @rawkode
    Helm
    Standardise Your Deployments
    

    View Slide

31. @rawkode
    APP_SECRET=67d829bf61dc5f87a73fd814e2c9f629
    framework:
    secret: %APP_SECRET%
    

    View Slide

32. @rawkode
    APP_SECRET=secret
    framework:
    secret: secret
    

    View Slide

33. @rawkode
    Terraform
    Can Provision K8s Primitives
    

    View Slide

34. @rawkode
    resource "random_string" "rpc-secret" {
    length = 32
    special = true
    }
    

    View Slide

35. @rawkode
    resource "kubernetes_secret" "service-vault" {
    data {
    rpc-secret = "${random_string.rpc-secret.result}"
    }
    }
    

    View Slide

36. @rawkode
    Fluentd
    Logging
    

    View Slide

37. @rawkode
    Fluentd
    Logging
    Written in
    C & Ruby
    

    View Slide

38. @rawkode
    Fluentd
    Logging
    13k events per second
    ~ 40MB RAM Footprint
    

    View Slide

39. @rawkode
    Fluent Bit
    Logging
    

    View Slide

40. @rawkode
    Fluent Bit
    Logging
    Written in
    C
    > 13k events per second
    ~ 450KIB RAM Footprint
    

    View Slide

41. @rawkode
    Fluentd/Bit
    Logging
    Plugins:
    ➔ AWS
    ➔ GCP
    ➔ MySQL / PostgreSQL
    ➔ Elasticsearch
    ➔ Docker
    ➔ Kubernetes
    ➔ Twitter
    ➔ Kafka
    

    View Slide

42. @rawkode
    Fluentd/Bit
    Docker Logging
    { “log-driver": "fluentd",
    "log-opts": {
    "fluentd-address": “...”
    }
    }
    

    View Slide

43. @rawkode
    Fluentd/Bit
    Kubernetes Logging
    # DaemonSet
    image:
    quay.io
    /fluent
    /fluentd-kubernetes-daemonset
    

    View Slide

44. @rawkode
    Prometheus
    Monitoring
    

    View Slide

45. @rawkode
    Prometheus
    Monitoring
    Written in
    Go
    

    View Slide

46. @rawkode
    Prometheus
    Monitoring
    Provides Dashboards
    &
    Works with Grafana
    

    View Slide

47. @rawkode
    Prometheus
    Monitoring
    Export Everything!
    ➔ Elasticsearch
    ➔ Consul
    ➔ MongoDB
    ➔ RabbitMQ
    ➔ Kafka
    ➔ Apache
    ➔ Nginx
    ➔ Fluentd
    ➔ …
    

    View Slide

48. @rawkode
    Prometheus
    Monitoring
    Scraping
    Magic with Kubernetes
    ➔ /metrics
    my_metric{label=a} 1
    my_metric{label=b} 6
    

    View Slide

49. @rawkode
    Prometheus
    Monitoring
    Manual Instrumentation
    $counter = newCounter([
    'namespace' => 'myApp',
    'subsystem' => 'Prod',
    'name' => 'httpRequests',
    ]);
    $counter->increment(
    ['url' => '/login',
    'status_code' => 200
    ], 1);
    

    View Slide

50. @rawkode
    Prometheus
    Monitoring
    Manual Instrumentation
    $counter = newCounter([
    'namespace' => 'myApp',
    'subsystem' => 'Prod',
    'name' => 'httpRequests',
    ]);
    $counter->increment(
    ['url' => '/login',
    'status_code' => 200
    ], 1);
    

    View Slide

51. @rawkode
    Prometheus
    Monitoring
    Manual Instrumentation
    $counter = newCounter([
    'namespace' => 'myApp',
    'subsystem' => 'Prod',
    'name' => 'httpRequests',
    ]);
    $counter->increment(
    ['url' => '/login',
    'status_code' => 200
    ], 1);
    

    View Slide

52. @rawkode
    OpenTracing
    Tracing
    

    View Slide

53. @rawkode
    OpenTracing
    Tracing
    A vendor-neutral open standard
    for distributed tracing
    

    View Slide

54. @rawkode
    OpenTracing
    Tracing
    Libraries available in 9
    languages:
    Go, JavaScript, Java, Python,
    Ruby, PHP, Objective-C, C++, C#
    

    View Slide

55. @rawkode
    OpenTracing
    Tracing
    Why?
    

    View Slide

56. @rawkode
    

    View Slide

57. @rawkode
    From Monolith to Cloud Native
    The 12 6-Step Programme
    

    View Slide

58. @rawkode
    1. Automation
    DevOps isn’t just a buzzword
    ➔ Environment Parity
    ➔ Continuous Integration
    ➔ Automated Tests
    ➔ Automated Deployment
    

    View Slide

59. @rawkode
    Trust
    

    View Slide

60. @rawkode
    2. 12-Factor
    Build Once. Deploy Anywhere.
    ➔ Version Control
    ➔ Explicit Dependencies
    ➔ JIT Configuration
    ➔ Build. Release. Run.
    ➔ Disposability
    

    View Slide

61. @rawkode
    Scalability
    Horizontally
    

    View Slide

62. @rawkode
    3. Adopt Containers
    Repeatable Deployments
    

    View Slide

63. @rawkode
    Idempotence
    

    View Slide

64. @rawkode
    4. Logging
    Understanding what is going
    wrong
    Centralised Logging
    ➔ ssh prod ☹
    ➔ Log to a central place
    ➔ Cross-sections
    ➔ Exception Tracking
    

    View Slide

65. @rawkode
    Intelligence
    

    View Slide

66. @rawkode
    5. Monitoring
    Predicting when something will go
    wrong
    Be the first to know when a
    service is down
    

    View Slide

67. @rawkode
    5. Monitoring
    Predicting when something will go
    wrong
    Understand your limits
    /
    Scale accordingly
    

    View Slide

68. @rawkode
    5. Monitoring
    Predicting when something will go
    wrong
    Don’t DDOS Yourself
    

    View Slide

69. @rawkode
    Intuition
    

    View Slide

70. @rawkode
    6. Micro-services
    Just don’t ask me to quantify
    “micro”
    ➔ Don’t refactor, replace
    ➔ Simplified testing
    ➔ Fast CI/Deploy
    ➔ Simple on-boarding
    

    View Slide

71. @rawkode
    Simplicity
    

    View Slide

72. @rawkode
    From Monolith to
    Micro-services
    ★ Scaleable
    Horizontally
    ★ Idempotence
    ★ Intelligence
    ★ Intuition
    ★ Trust
    ★ Simplicity
    

    View Slide

73. @rawkode
    

    View Slide

74. @rawkode
    Service Mesh
    (istio edition)
    Connect, Manage & Secure your
    services
    ➔ Load Balancing
    ➔ Canaries
    ➔ Circuit Breakers
    ➔ Handling Timeouts and Retries
    ➔ Fault Injection
    ➔ mTLS
    

    View Slide

75. @rawkode
    Service Mesh
    (istio edition)
    Load Balancing
    source:
    name: user-service
    labels:
    version: v3
    destination:
    name: email-service
    labels:
    version: v1
    loadBalancing:
    name: ROUND_ROBIN
    

    View Slide

76. @rawkode
    Service Mesh
    (istio edition)
    Load Balancing
    Canaries
    destination:
    name: user-service
    route:
    - labels:
    version: v2
    weight: 25
    - labels:
    version: v1
    weight: 75
    

    View Slide

77. @rawkode
    Service Mesh
    (istio edition)
    Circuit Breakers
    circuitBreaker:
    simpleCb:
    maxConnections: 100
    

    View Slide

78. @rawkode
    Service Mesh
    (istio edition)
    Timeouts
    &
    Retries
    httpReqTimeout:
    simpleTimeout:
    timeout: 10s
    httpReqRetries:
    simpleRetry:
    attempts: 3
    

    View Slide

79. @rawkode
    Service Mesh
    (istio edition)
    Fault Injection
    httpFault:
    delay:
    percent: 10
    fixedDelay: 5s
    abort:
    percent: 2
    httpStatus: 400
    

    View Slide

80. @rawkode
    But wait, there’s more!
    

    View Slide

81. @rawkode
    OpenTracing
    Automatic Tracing of gRPC & HTTP
    requests
    

    View Slide

82. @rawkode
    mTLS
    Automatic mTLS between service
    requests
    

    View Slide

83. @rawkode
    When will I know I’m there?
    

    View Slide

84. @rawkode
    

    View Slide

85. @rawkode
    You are now Cloud Native
    

    View Slide