KubeCon Keynote 2016- Distributed Systems Simplified on Kubernetes

Demo https://github.com/philips/etcd-prometheus-operator-demo

Transcript

1. Demo https://github.com/philips/etcd-prometheus-operator-demo
2. Brandon Philips @brandonphilips | brandon@coreos.com | coreos.com Distributed Systems Simplified Using Kubernetes
3. Overview Scaling Stateless Apps
4. ReplicaSet
5. ReplicaSet
6. Overview Scaling Complex Apps
7. $ kubectl run db --image=quay.io/my/db Creating a Database is Easy on Kubernetes
8. • Resize/Upgrade - coordination for availability • Reconfigure - tedious generation / templating • Backup - requires coordination on instances • Healing - observe and act for recovery Managing a Distributed Database is Harder
9. Idea: Extend Kubernetes
10. $ cat my-db-cluster.yaml spec: clusterSize: 3 readReplicas: 2 version: v4.0.1 The Dream
11. Introducing Operators
13. etcd Overview • Introduced in 2013 by CoreOS • Primary datastore of Kubernetes • Auto-leader election for availability
14. $ cat etcd-cluster.yaml spec: clusterSize: 3 version: v3.1.0 etcd Operator Resource
15. etcd Operator
16. etcd Operator Open-source github.com/coreos/etcd-operator
17. Kubernetes self-hosting etcd Easy HA Setups on Kubernetes Automated backup to object store Chaos Monkey Testing for 1.0 etcd Operator - Future Work
18. Prometheus Operator • Operates Prometheus on k8s • Handles common tasks: ◦ Create/Destroy ◦ Monitor Configuration ◦ Services Targets via Labels • Configured by resources
19. Prometheus Operator Also Open-source github.com/coreos/prometheus-operator
20. • Read more at coreos.com/blog • Try out these new Operators • Build and Discuss other Operators (redis, postgres, etc) Next Steps
21. Thank you! Brandon Philips @brandonphilips | brandon@coreos.com | tectonic.com We’re hiring in all departments! Email: careers@coreos.com Positions: coreos.com/ careers