The road to Kubernetes

Come and learn about the history of Kubernetes from Google, and a little about the future of the platform

Transcript

1. DevOps with Kubernetes Craig Box Cloud Advocate Deploying and managing microservices
2. Deploying Software
4. One application, many users
8. DevOps Operators Developers
9. DevOps Procure Deploy Secure Monitor Schedule Isolate Control access Update Roll back Remediate Troubleshoot Scale Write code
10. "As a developer, I considered Operations skills akin to superpowers. Writing code that worked on my laptop was easy — getting it to work across a cluster of machines was orders of magnitudes harder." Cindy Sridharan, imgix “ “
11. Your laptop
12. Production
13. • A set of tasks • Relating to one or more applications • Continuously delivered • On one machine? • On many machines? • On one remote provider/cloud? • On many clouds? Operations
14. Operators Developers DevOps Operators Developers Contract
15. • Size teams correctly • Scale independently • Decouple applications too • "Breaking up the monolith" Hardware Specialisation Network SRE Storage SRE Cluster SRE App SRE Developers
16. Number of running containers Core Ops Team Enabled Google to grow our fleet over 10x faster than we grew our ops team
18. Container-centric infrastructure for deploying and managing applications. What is Kubernetes? “ “
19. What is a container? An application bundle A process group with some restrictions VM: Starting an OS Booting phone: ~60 seconds Changeable data: mixed with image Container: Starting a process Tap an app: ~0.2 seconds Changeable data: stored elsewhere
20. Containers Docker rkt OCI
21. Service Controller Pod Kubernetes components A loop that drives current state towards desired state Example: ReplicaSet A small group of tightly coupled containers Example: Static site generator & web server A set of running pods that work together Example: Load-balanced backends
22. Description of what you want to run You don't say where you want it to run: Kubernetes figures that out for you Manifests apiVersion: v1 kind: Pod metadata: name: redis-django labels: app: webapp spec: containers: - name: key-value-store image: redis ports: - containerPort: 6379 - name: frontend image: django ports: - containerPort: 8000
23. • Scheduling • Lifecycle and health • Scaling • Naming and discovery • Load balancing • Storage volumes • Logging and monitoring • Debugging and introspection • Identity and authorization • Piranha proof infrastructure Kubernetes provides
24. The new contracts for distributed systems Developer "Operating system" "Hardware" Contract Contract
25. A simple, fast, and fully-managed way to run containerized applications, from test to production. Developers
26. Developer time is at a premium. Kubernetes ease of deployment accelerates everything. Developers go from commit to deploy in production in less than 10 minutes allowing us to deploy 30-50 times a day Tyler Davis, DevOps “ ”
27. A standardized platform for reliable and scalable clusters without operational overhead. Operators
28. The industry-leading open source infrastructure platform, enabling interoperability, removing vendor lock-in, and providing cost savings. CTOs
29. Consistency everywhere On-Premise Cloud Hybrid
30. While Kubernetes did nudge us in the direction of GCP, we’re more cloud agnostic than ever because Kubernetes can live anywhere. Bas Moeys, co-founder & Head of Technology “ ”
32. Cloud Native
35. Container packaged Predictable deployment Efficient resource isolation Dynamically scheduled Higher QoS and efficiency Lower ops cost Microservices oriented Domain driven Independently upgradeable Cloud Native
37. Community composition Before 1.0 GA Between 1.0 - 1.5
38. Community composition Welcome!
42. "The dial tone of the Internet"
43. Let Google be your SRE
45. Fully managed Kubernetes
46. Google manages your control plane Container Engine Kubernetes Master API Server Controller Manager Scheduler etcd • Backups • Monitoring • Restarts • Resizing for larger clusters • 99.5% SLA
47. ...and system components on your nodes Container Engine Kubernetes Master API Server Controller Manager Scheduler etcd Container Engine Kubernetes Nodes Logging Monitoring Ingress backend Runtimes
48. Node management features Node upgrade: • Update Kubernetes version • Update node OS Node repair: • Automatically repair broken nodes Container Engine Kubernetes Nodes Logging Monitoring Ingress backend Runtimes
50. Full audit trail List operations Describe an operation to get more details $ gcloud container operations list NAME TYPE .. STATUS operation-15-0a UPGRADE_MASTER .. DONE operation-11-78 AUTO_UPGRADE_NODES .. DONE $ gcloud container operations describe <op> name: operation-15-0a operationType: UPGRADE_MASTER selfLink: ... status: DONE targetLink: ... zone: us-west1-b
52. A sneak peek
53. The shift to a modern distributed architecture has left enterprises unable to monitor, manage or secure their services in a consistent way.
54. What is a service mesh? A network for services, not bytes • Visibility • Resiliency & Efficiency • Traffic Control • Security • Policy Enforcement
55. HTTP/1.1, HTTP/2, gRPC, TCP with or without TLS HTTP/1.1, HTTP/2, gRPC, TCP with or without TLS Internet A service mesh
57. The Istio service mesh • A complete framework for connecting, securing, managing and monitoring services • Secure and monitor traffic for microservices and legacy services • An open platform with key contributions from Google, IBM, Lyft and others • Multi-environment and multi-platform
58. • Connect Resiliency, discovery, load balancing • Manage Traffic control, policy enforcement • Monitor Metrics, Logging, Tracing • Secure End-to-end Authentication and Authorization The Istio service mesh
59. • kubernetes.io • cloud.google.com/container-engine • istio.io Learn more
60. Thank you @craigbox #gdgspb