The Hitchhiker’s guide to the cloud

A colorful introduction to modern cloud solutions.
From Docker to Helm and what is a Kubernetes anyway?

Transcript

1. The Hitchhiker’s guide to the cloud
3. Do you remember just coding in your favorite language?
16. I have been there too
17. Let me introduce myself: Arno Koehler Engineering Manager iO Developer since 2005 JVM Developer since 2006
18. Software is creativity
19. In the 80’s these guys found out that 10x developers just had more fun and focus
20. API Migration At Schiphol
21. Documentation is a game changer
22. Augmented Development with Ai your new Rubber Ducky
23. Architecture Decision Records (ADR)
24. Let us start our journey The road to understanding all of it
25. Docker Containers vs Virtual Machines
26. Source: https://www.trianz.com/insights/containerization-vs-virtualization
27. Source: https://diadem.in/blog/docker-vs-vm-hypervisor-vs-docker/
28. How to create a Docker container?
29. Dockerfile Dockerfile Dockerfile
30. Frontend Baseimages Java Baseimages Java images are correct Node Images may contain hallucinations
31. `FROM scratch` WHY?! • Smallest possible image size (can be <10MB for compiled applications) • Improved security (no unnecessary components) • Minimal attack surface • Faster startup and deployment • Reduced network transfer times in CI/CD pipelines • No shell, no package manager, no utilities - only your compiled binary
32. Docker is cool We now know how to setup a Docker Container
33. Docker-compose
34. Docker Compose can do - restarts - network segmentation On a single NODE
35. Docker compose yml
36. Docker-compose is cool Only for a single node so nice for simple local development
37. Kubernetes
38. Kubernetes
39. Clusters
40. Nodes
41. POD
42. And another POD
43. Clusters
44. Clusters
45. API SERVER Scheduler Controller Manager etcd Control plane
46. Control plane variants
47. Controller Or Control Loop
48. Controllers Control loop
51. Namespaces
52. Back to the POD
53. POD .. or this
54. What is a POD?
56. StatefulSets
57. StatefulSets
58. ConfigMaps
59. Secrets
60. Deployment
61. https://github.com/iodigital-com/kubernetes-greenblue-workshop
63. Registry
65. Services
66. Service Types
68. Vertical Scaling
69. Horizontal Scaling
71. Reverse Proxy
72. Castle Moat Ingress
73. ISTIO
74. Kubernetes
75. Learning on your machine
77. `docker build`
78. Local registry
80. https://github.com/iodigital-com/kubernetes-greenblue-workshop
82. Deploying something in Kubernetes: kubectl apply -f blue-deployment.yaml kubectl apply -f service-node-port-blue.yaml
83. Kubectl describe pod <pod-name> Kubectl logs Kubectl –l app=my-app,version=blue Flags to know: -f follow the logs -p previous pod (current pod-1) -c specific container within a pod only
86. TOOLS
87. Green Blue deployments
88. Ingress to switch
89. Canary Releases
90. Canary Ingress
91. Kubernetes
93. Kustomize || HELM
97. HELM ≈ Brew
99. Public Helm Charts
100. Overriding values
102. Default structure HELM project
103. Chart.yaml
104. Chart.yaml
105. Overriding values from Kubernetes
106. Flags you can use
107. Since Helm 3 They removed TILLER Important note: TILLER == Controller
108. Upgrade & Reuse values
109. How to look at the past
110. More flags to know
112. values.yaml Values.yaml
113. Templates sandbox.yaml Go template
114. Render without deployment
115. Linting
116. Route Add some fun
118. A real life HELM repo
119. Deployment
120. Service
121. • Helm upgrade —install Please remember this
123. GITLAB
126. Questions Key Challenges in the Current Landscape The cloud tooling space faces several significant challenges: 1.Complexity Overload: 1.The sheer number of tools and approaches creates cognitive overload for teams. 2.From my experience, it's evident that there’s an expansion in tools trying to solve similar problems in slightly different ways. 2.Integration Challenges: 1.Many tools solve specific problems well but don't necessarily integrate seamlessly with other parts of the toolchain. 3.Learning Curve: The learning curve for many cloud-native tools creates barriers to adoption and effective use. 4.Configuration Management: Managing configurations across multiple environments and at scale remains difficult, which is why tools like Kluctl are emerging to address this specific pain point.
127. Questions How do you see the future? More tools are coming to the market to bridge certain gaps. Tools like ArgoCD/Flux and Kluctl are trying to fill different gaps and more of these tools will probably be written in the future. When choosing a tool, consider the complexity of your deployment, existing investments and your team expertise.