Deliver cloud-native microservices applications with Zadig

微服务示例

Here we use the open source project is https://github.com/GoogleCloudPlatform/microservices-demo,The open source project called Online Boutique(https://onlineboutique.dev/),Is a native of cloud services demo application,其中包含 11 个微服务,The application is a based on Web The e-commerce application,In which users can browse the goods、Add them to the cart and buy them.

Based on the project for Zadig Optimized modification,项目地址： https://github.com/cnych/microservices-demo

该项目由 11 A micro services written in different languages,它们通过 gRPC 相互通信.架构如下所示.

微服务架构

The function of each service description shown in the following table.

服务功能

Zadig 项目

接下来我们来使用 Zadig To delivery of the services,Use front and similar.

首先新建一个名为 microservices-demo 的项目,项目类型为 K8s YAML 项目.

新建项目

Click the new button immediately,And then into the project initialization page.

项目初始化

Click next to enter service configuration page,We know that service template can manually create、Code synchronization in warehouse or existing K8s Resources in the import and to.We here in service template code warehouse,So choose to synchronize from the code base,选择对应的代码仓库、Branch and the corresponding resource directory listing,Then click the sync button to synchronize the corresponding service to Zadig 中来.

同步服务

We can see contains after synchronization to come over here 12 个服务,Each service template were directly display the,But because it is through the warehouse synchronous,A template is a read-only mode,On the right side will be read automatically to the corresponding image information service.

注意：Zadig After reading the resource list, 会以 K8s YAML The container name as the only key 进行去重 ,所以在编写 K8s YAML Don't let the container name repeat,Otherwise the import will be lost after service.

服务模板

Click on the right side of the read behind the image of the service component of add build button,To configure the service image of how to build.

First of all need to add the service code source information,The code here in GitLab 上面,So to add the corresponding code warehouse and branch information.We here all service code in the code for the warehouse under the root directory of src 目录下.

代码结构

这属于典型的 Monorepo 类型的仓库（单体）,Each service and we are not configured to submodule,So all throughout the code to each service the build warehouse Clone 下来,But actually we only need one of these services can be,Git Is to support the operation of,For example, we now just want to get adservice The service data,Could be obtained by the following way.

$ mkdir microservices-demo && cd microservices-demo
$ git init
$ git remote add origin [email protected]:course/microservices-demo.git
$ git config core.sparsecheckout true
$ git sparse-checkout set "src/adservice"
$ git pull --depth 1 origin main
remote: Enumerating objects: 307, done.
remote: Counting objects: 100% (307/307), done.
remote: Compressing objects: 100% (234/234), done.
remote: Total 307 (delta 72), reused 185 (delta 39), pack-reused 0
Receiving objects: 100% (307/307), 9.55 MiB | 5.28 MiB/s, done.
Resolving deltas: 100% (72/72), done.
From git.k8s.local:course/microservices-demo
 * branch            main       -> FETCH_HEAD
 * [new branch]      main       -> origin/main
$ ls -la src
total 0
drwxr-xr-x   3 cnych  staff   96 Jul 15 14:10 .
drwxr-xr-x   4 cnych  staff  128 Jul 15 14:10 ..
drwxr-xr-x  12 cnych  staff  384 Jul 15 14:10 adservice

Through the above ways can only obtain the specified directory code,但是遗憾的是 Zadig This function does not currently support,Subsequent may support！

After the code source configuration,The most important is to add general build script.

构建服务

We have here is actually configure how to build the mirror,The corresponding script as shown below：

#!/bin/bash
set -e
cd $WORKSPACE/$REPONAME_0/src/adservice
docker build -t $IMAGE -f Dockerfile .
docker push $IMAGE

First you need to enter to the current service under the root directory of the code,我们这里使用的是 cd $WORKSPACE/$REPONAME_0/src/adservice 命令,其中的 $WORKSPACE、$REPONAME_0、$IMAGE Both for the construction of the built-in variables,$WORKSPACE Said work root directory,而 $REPONAME_0 The name of the first code warehouse said configuration,也就是 microservices-demo.

构建变量

After entering into the service under the root directory,我们只需要执行 docker build Command to build the mirror can be,Each service is provided for under the root directory of the Dockerfile 文件,Build image using variable $IMAGE 代替,Will use add the default image of warehouse.

The default image of naming rules as shown in the following,We can also custom.

The mirror command rules

After finish build configuration,Remember to save building,Configuration in the same way all service build.

After the completion of the service configuration start next to join the environment,Also by default will automatically create a set of dev 和 qa 的环境以及 3 条工作流.

加入环境

After joining the environment can see the corresponding 3 The workflow,点击完成即可.

工作流

执行工作流

这样项目就创建成功了,Now we are in dev Environment to run the workflow.点击执行工作流,Can choose what we want to build service in service,You can select a can also select multiple services.

执行任务

After the same task execution will perform the configuration general-purpose scripting,And then the corresponding service deployment to K8s 集群中去.

任务详情

Will carry out in the environment after the completion of all services page can see all of the service status and the latest image information.

dev环境

Each service deployed to properly dev 环境后,查看对应的 Pod 状态：

$ kubectl get pods -n microservices-demo-env-dev
NAME                                     READY   STATUS    RESTARTS   AGE
adservice-5b5b97cf59-b7d6g               1/1     Running   0          21m
cartservice-7d66bd5c4-wxd6v              1/1     Running   0          15m
checkoutservice-6dbc8cdc46-fqrnx         1/1     Running   0          14m
currencyservice-54fccf7b9f-c29h5         1/1     Running   0          13m
emailservice-844b6c9b58-j5smc            1/1     Running   0          13m
frontend-649699cc5-s5ggp                 1/1     Running   0          12m
loadgenerator-565cdbb5dc-6dtrb           1/1     Running   0          5m42s
paymentservice-fd8f95f79-4889n           1/1     Running   0          11m
productcatalogservice-775db476b6-l4s2t   1/1     Running   0          6m47s
recommendationservice-79f558bc9b-jx69n   1/1     Running   0          6m47s
redis-cart-f9bdd7959-qtc9n               1/1     Running   0          34m
shippingservice-f789c4494-ktbnz          1/1     Running   0          6m43s
$ kubectl get svc -n microservices-demo-env-dev
NAME                    TYPE           CLUSTER-IP       EXTERNAL-IP    PORT(S)        AGE
adservice               ClusterIP      10.96.40.23      <none>         9555/TCP       41m
cartservice             ClusterIP      10.110.0.136     <none>         7070/TCP       41m
checkoutservice         ClusterIP      10.109.209.29    <none>         5050/TCP       41m
currencyservice         ClusterIP      10.98.43.229     <none>         7000/TCP       41m
emailservice            ClusterIP      10.101.67.47     <none>         5000/TCP       41m
frontend                ClusterIP      10.96.37.222     <none>         80/TCP         41m
frontend-external       LoadBalancer   10.109.168.181   192.168.0.53   80:30090/TCP   41m
paymentservice          ClusterIP      10.105.223.48    <none>         50051/TCP      41m
productcatalogservice   ClusterIP      10.102.112.64    <none>         3550/TCP       41m
recommendationservice   ClusterIP      10.99.127.143    <none>         8080/TCP       41m
redis-cart              ClusterIP      10.111.15.205    <none>         6379/TCP       41m
shippingservice         ClusterIP      10.104.122.207   <none>         50051/TCP      41m

The last is through the service frontend-external 这个 LoadBalancer 类型的 Service To expose the service,We can directly through the distribution of IP 192.168.0.53 来访问服务了.

frontend external

触发器

We can also give same service add a trigger to trigger the task,In the workflow editor page click on the left side of the trigger to add,勾选 Webhook -> 添加配置,We can add a only for first adservice The service of the trigger,Select the corresponding code library、To deploy the service,File directory is the most critical part of the configuration,What is the code in the warehouse file changes would trigger our mission,我们的配置为：

src/adservice/
!.md
!.gitignore

The representation of the segment configuration when code in the warehouse src/adservice/ Folder has the following code changes,并且不是 .md 或者 .gitignore File will trigger the task.

触发器

Remember to keep the configuration when.Now we can go to modify the adservice 服务中的代码,修改代码 src/adservice/src/main/java/hipstershop/AdService.java,比如我们将 177 行的 50%修改为 60%,然后提交代码到 main 分支.

修改代码

When we refer to the above code push 到 main 分支后,Zadig Immediately triggered a new task.

webhook trigger

The task execution after the success we can go to see if the product's page effect.

验证修改

Here I will realize the micro services continue to build,Also we can go to manually create an environment for delivery,Operating the same way.

测试用例

In addition we can also to create new test cases,In the test page click the test button to create new test cases.

新建测试

For example, we here for a go Service to do a simple test,In dependent packages can choose corresponding rely on,If there is no corresponding packages,则可以新建一个.

软件依赖

比如我们需要一个 1.17 版本的 go 环境,可以通过 系统设置 -> 软件包管理 新建一个软件包.

新建软件包

And then we back to create a test case,选择相应的依赖,The same configuration corresponding source code and test script.

测试脚本

After creation can be in the same way to perform this test case.

执行测试

Test details page and the workflow task basically consistent.

测试详情

Also can see the corresponding test report.

测试报告

In addition we can also be automated testing and workflow associated,After the update daily operation workflow environment,会自动执行自动化测试.Can be achieved as long as the environment changes,Is the first time to do the test automation.

Associated workflow

Associated after start the workflow task can see the test case.

测试用例

工作流

Zadig Will also help us build data statistics,Provide building efficiency、Deployment efficiency data such as.

Building efficiency

到这里我们就完成了使用 Zadig Continuous integration and micro service project for delivery,Of course in the actual production conditions and specific project business relationship,This needs to be able to in combination with actual needs to practice.