Barriers to adoption of Kubernetes in production

https://www.youtube.com/watch?v=Vw62jhsZxU8
https://github.com/weaveworks/kubernetes-anywhere

- motivation
- barriers to adoption
- our team’s experience
- tools/demo

Transcript

1. if you can drink and watch Star Wars, you can handle Kubernetes in production (because that’s exactly what we did) Ilya Dmitrichenko
2. 1. motivation 2. barriers to adoption 3. our team’s experience 4. tools/demo kube-1 kube-3 kube-2 kube-4 kube-6 kube-5 kube-7
3. deliver faster your organisation absolutely must
4. adopt microservices everyone in your team agreed that you will
5. adopt containers everyone in your team agreed that you will
6. adopt kubernetes everyone in your team agreed that you will
7. faster development everyone in your team sees that microservices are key to
8. reliable deployment everyone in your team sees that containers are key to
9. scalability & utilisation everyone in your team sees that kubernetes is key to
10. and most of people in this room will agree with you.
11. and most of people in this room will agree with you. in theory.
12. and most of people in this room will agree with you. in theory. while in practice, there are still more questions…
13. more questions will arise because
14. more questions will arise because your team has opinions
15. more questions will arise because your team has opinions you cannot take very big risk
16. more questions will arise because your team has opinions you cannot take very big risk and legacy systems are a thing.
17. let’s see what these questions are…
18. deliver faster don’t forget that you must
19. 1. language & frameworks 1/5
20. 1. language & frameworks you know what’s best for your application
21. 1. language & frameworks with microservices this is much easier then it used to be
22. 2. toolchains & workflow 2/5
23. 2. toolchains & workflow you already know how to compile your app, write tests & configure CI
24. 2. toolchains & workflow building containers is easy, most major CI systems support it
25. 3. infrastructure 3/5
26. 3. infrastructure creating a Kubernetes cluster is fairly easy, for some users
27. 3. infrastructure not as easy if you have opinions and non-trivial requirements
28. 4. shipping pipeline 4/5
29. 4. shipping pipeline Kubernetes has all the primitives, but tooling isn’t fully understood yet
30. 4. shipping pipeline what is the best practice to handle deployment of your application?
31. 5. operations 5/5
32. 5. operations what are the tools you need to execute most efficiently?
33. let’s summarise
34. we call this ABCDE of Microservices A) write the app B) test and build it C) provision a cluster D) deploy the app E) execute
35. deliver faster? that’s all theory, but how are you going to
36. practical concerns about infrastructure
37. practical concerns about infrastructure opinions & risks
38. practical concerns about infrastructure all-new and shiny
39. operating system common examples of infrastructure opinions
40. automation tools common examples of infrastructure opinions
41. common examples of infrastructure opinions cloud provider
42. common examples of infrastructure opinions virtual vs bare-metal
43. common examples of infrastructure opinions commercial support
44. common examples of infrastructure opinions cluster management
45. common examples of infrastructure opinions troubleshooting
46. common examples of infrastructure opinions databases
47. common examples of infrastructure opinions access control
48. common examples of infrastructure opinions platform as a service?
49. kubernetes depends on infra a lot all of the opinions will surface
50. Controller Manager [kube-controller-manager] Scheduler [kube-scheduler] API Server [kube-apiserver] Cluster Nodes [kube-proxy] [kubelet] etcd1 etcd2 etcd3 all of the opinions will surface and you still need really?
51. Controller Manager [kube-controller-manager] Scheduler [kube-scheduler] API Server [kube-apiserver] Cluster Nodes [kube-proxy] [kubelet] etcd1 etcd2 etcd3
52. Controller Manager [kube-controller-manager] Scheduler [kube-scheduler] API Server [kube-apiserver] Cluster Nodes [kube-proxy] [kubelet] etcd1 etcd2 etcd3 all of the opinions will surface and you still need a network
53. Controller Manager [kube-controller-manager] Scheduler [kube-scheduler] API Server [kube-apiserver] Cluster Nodes [kube-proxy] [kubelet] etcd1 etcd2 etcd3 all of the opinions will surface and you still need a storage product
54. Controller Manager [kube-controller-manager] Scheduler [kube-scheduler] API Server [kube-apiserver] Cluster Nodes [kube-proxy] [kubelet] etcd1 etcd2 etcd3 all of the opinions will surface and you still need a monitoring solution
55. Controller Manager [kube-controller-manager] Scheduler [kube-scheduler] API Server [kube-apiserver] Cluster Nodes [kube-proxy] [kubelet] etcd1 etcd2 etcd3 all of the opinions will surface and you still need a private registry
56. Controller Manager [kube-controller-manager] Scheduler [kube-scheduler] API Server [kube-apiserver] Cluster Nodes [kube-proxy] [kubelet] etcd1 etcd2 etcd3 all of the opinions will surface and you still need to deliver faster
57. our team’s use-case scope-app-259 0 pods app-mapper 2 pods The Internet users 2 pods kubernetes 0 pods Uncontained ip-172-20-0-149 Uncontained ip-172-20-0-148 scope-app-327 1 pod scope-app-281 0 pods frontend 2 pods scope-app-301 0 pods fairly simple Go app, only 6 services
58. our team’s use-case scope-app-259 0 pods app-mapper 2 pods The Internet users 2 pods kubernetes 0 pods Uncontained ip-172-20-0-149 Uncontained ip-172-20-0-148 scope-app-327 1 pod scope-app-281 0 pods frontend 2 pods scope-app-301 0 pods no messaging or fancy databases
59. our team’s use-case scope-app-259 0 pods app-mapper 2 pods The Internet users 2 pods kubernetes 0 pods Uncontained ip-172-20-0-149 Uncontained ip-172-20-0-148 scope-app-327 1 pod scope-app-281 0 pods frontend 2 pods scope-app-301 0 pods must have portable cluster deployment
60. our team’s experience
61. our team’s experience tried a few different approaches for local and cloud deployments
62. our team’s experience some cloud providers are less equal than others
63. our team’s experience some cloud provider services are slower than others
64. our team’s experience setup from scratch is hard, especially due to enforced security
65. our team’s experience self-hosting is nice, but somewhat harder to reason about
66. our team’s experience operating systems and and configuration
67. our team’s experience declarative provisioning wins and network-first was a revelation
68. our team’s experience checkout git.io/k8s-anywhere and see for yourself
69. we needed simple and robust provisioning in any environment
70. we needed zero-config scale-out and in-place upgrade
71. we built easy to adopt cluster provisioning tools
72. we built containerised public key infrastructure solution
73. we built weave as a bootstrap/management and pod networking
74. in a nutshell weavedns, few small shell scripts and hyperkube images
75. in a nutshell just pick a few docker hosts and launch weave
76. in a nutshell pick worker, master & etcd nodes. and off you go!
77. live demo ip-172-20-0-164 ip-172-20-0-222 ip-172-20-0-244 ip-172-20-0-243 ip-172-20-0-242 ip-172-20-0-12 ip-172-20-0-51
78. potentially. are you anywhere closer to delivering faster?
79. scope.weave.works ilya-proxy-e20b5 1 container metrics-98d4fd7-ybopj 1 container dbshell-fons-e6791 1 container users-cb08840-kjd39 1 container dbshell-fons-14417 dbshell-fons-2eab5 dbshell-fons-2d95f Uncontained ip-172-20-0-149 Uncontained ip-172-20-0-148 app-mapper-77d895a-rjqc1 app-mapper-77d895a-bk4k5 1 container The Internet users-cb08840-dj03q 1 container frontend-515708e-0y9h0 1 container
80. @errordeveloper
81. weave release v1.5
82. party! users-cb08840-kjd39 2 containers metrics-98d4fd7-ybopj 2 containers users-cb08840-dj03q 2 containers frontend-515708e-0y9h0 2 containers dbshell-fons-e6791 2 containers app-mapper-77d895a-bk4k5 2 containers Uncontained ip-172-20-0-149 ilya-proxy-e20b5 2 containers The Internet Uncontained ip-172-20-0-148