让开发人员和运维人员发挥各自所长

Kubernetes 为应用程序开发人员和运维人员提供了陡峭的学习曲线,以帮助其学习管理多个工具和工件以构建部署管道的新概念。在云和混合云环境中,组件不断升级、扩展和销毁,这样的动态性和异构性使得此问题进一步加剧。 在此次演讲中,我们想要展示通过分离开发和运维工件、并在单独的存储库中对它们进行版本控制的工作流程,以使开发人员和运维人员发挥各自所长。这种分离操作可以使 CI/CD 工具有效保持理想状态与实际状态同步,同时简化交付多种应用时所需的复杂过程。最后,我们将展示在管理不同环境中的多个 Kubernetes 集群时,如何使用 OSS Cluster Registry 管理部署和实验过程
展开查看详情

1.Let Developers and Operators Focus on What They Know Best Rafal Gajdulewicz, Software Engineer Julie Zhuying Li, Interaction Designer

2. developers - writing code - testing code - observability PRODUCT DEVELOPMENT Core tasks

3. developers - writing code - testing code - observability PRODUCT DEVELOPMENT Core tasks operators - test application - monitoring - upgrades, updates

4. devs General development and release workflow request deployment code commit to test

5. devs ops General development and release workflow request deployment what to build and code commit to test deploy, where publish test deployment done, deploy artifact let devs know to test env

6. devs ops General development and release workflow request deployment what to build and code commit to test deploy, where publish test request deployment verify deployment deployment done, deploy artifact to prod and correctness let devs know to test env

7. devs ops General development and release workflow request deployment what to build and code commit to test deploy, where publish test request deployment verify deployment deployment done, deploy artifact to prod and correctness let devs know to test env what to build and deploy artifact shipped! deploy, where publish prod to prod env

8. devs ops Can we automate some of these code commit request deployment what to build and to test deploy, where publish test steps so that developers and operators can focus on what they request deployment verify deployment deployment done, deploy artifact to prod and correctness let devs know to test env know best ? what to build and deploy artifact ship it deploy, where publish prod to prod env

9.

10. Let’s use some tools that are familiar version control CI/CD cluster registry CLUSTER REGISTRY AND MORE ...

11. CI/CD Tools automated steps Where CI/CD tools can help us request deployment what to build and code commit to test deploy, where publish test request deployment verify deployment deployment done, deploy artifact to prod and correctness let devs know to test env what to build and deploy artifact shipped! deploy, where publish prod to prod env

12. cluster registry Where cluster registry can come in request deployment what to build and code commit to test deploy, where publish test request deployment verify deployment deployment done, deploy artifact to prod and correctness let devs know to test env what to build and deploy artifact shipped! deploy, where publish prod to prod env

13. devs ops And you should probably avoid doing this request deployment what to build and code commit to test deploy, where publish test request deployment verify deployment deployment done, deploy artifact to prod and correctness let devs know to test env what to build and deploy artifact shipped! deploy, where publish prod to prod env

14. devs ops After these jobs automated, your pipeline gets much shorter request deployment what to build and code commit to test deploy, where publish test git, CI, cluster registry deploy to test request deployment verify deployment deployment done, deploy artifact to prod and correctness let devs know to test env deploy the artifact build what to and to canary, observedeploy artifact metrics / shipped! deploy, probers, decide ifpublish where rolloutprod to prod env should be continued approval

15. After optimization devs ops deploy the git, CI, CR request artifact to verify deployment code commit deploy to deployment to prod env and correctness test prod approval shipped!

16.Introduction Demo

17.Before devs = 5 steps ops = 7 steps After devs = 4 steps ops = 1 steps

18.Introduction Thank you