rollout-patterns-smoothly-migrating-and-rolling-out-your-microservices-zhu-ya-daepkui-hui-zhe-rezhu-ya-zha-tim-xiao-daocloud.pdf

1、Rollout Patterns:Smoothly Migrate and Roll Out Your MicroservicesTim Xiao,DaoCloudAbout meTim muma378DaoCloud,Software Engineer,DevOps Platform LeaderArgo Project ContributorHiking、BadmintonandCoffeeLoverBackgroundServices were built with Java and Dubbo2Microservices architectureMigrate to service m

2、eshWant canary rolloutIstioEnvoyDubboProviderpodEnvoyDubboConsumerpodRollout PatternsPATTERN 1:One service at a timeUsually happen for monolithic or compatible interfacesThe delivery order could be very importantLow requirement on Ops but high requirement on DevHow you deliver your services?Rollout

3、PatternsThe ideal scenario for argo-rolloutsWhat is Argo-Rollouts(optionally)integrates with ingress controllers and service meshes,leveraging their traffic shaping abilities to gradually shift traffic to the new version during an updateWhat is Argo-Rolloutscan query and interpret metrics from vario

4、us providers to verify key KPIs and drive automated promotion or rollback during an updatePattern 1HoweverWhen a change involves multiple components,they are preferred to be delivered in togetherNot only will the workloads be updated,but the configuration policies will also changeRollout PatternsPAT

5、TERN 2:multiple services,each is backward-compatibleCompared with pattern 1,applications are delivered as a wholeBackward-compatible is for callers,not for calleesTraffic become very complicate,hard to troubleshoot,hard to rollbackerror pronePattern 2All you need is a delivery tool to control the or

6、der of deployingWhat is Argo CDArgo CD follows the GitOps pattern of using Git repositories as the source of truth for defining the desired application stateArgo CD automates the deployment of the desired application states in the specified target environmentsArgo CD reports&visualizes the differenc