Netflix Conductor 入门Example

通过命令行将Netflix Conductor Sever端启动之后（ https://netflix.github.io/conductor/intro/#installing-and-running 介绍了如何安装Conductor），访问localhost:8080/swagger-ui.html地址显示如下页面：

swagger.png

访问localhost:5000地址显示的是ui页面

ui.png

我们假设有一个流程，该流程根据输入的城市名去查询该城市的天气，如果气温大于37度则发送一条短信通知指定的人。

定义任务

首先我们要定义一个查询天气的任务和一个发送短信的任务。通过postman或者swagger向conductor server提交这两个任务。代码如下

curl --location --request POST 'http://localhost:8080/api/metadata/taskdefs' \
--header 'Content-Type: application/json' \
--data-raw '[
{
  "name": "queryWeather",
  "inputKeys": [
    "city"
  ],
  "outputKeys": [
    "temperature"
  ],
  "retryCount": 3,
  "retryLogic": "FIXED",
  "retryDelaySeconds": 10,
  "timeoutSeconds": 300,
  "timeoutPolicy": "TIME_OUT_WF",
  "responseTimeoutSeconds": 180,
  "ownerEmail": "[email protected]"  
}
,
{
  "name": "sendMessage",
  "inputKeys": [
    "receiver",
    "content"
  ],  
  "retryCount": 3,
  "retryLogic": "FIXED",
  "retryDelaySeconds": 10,
  "timeoutSeconds": 300,
  "timeoutPolicy": "TIME_OUT_WF",
  "responseTimeoutSeconds": 180,
  "ownerEmail": "[email protected]"  
}
]'

提交成功后，可以在localhost:5000/taskDef页面看到

taskDef.png

编排任务

定义完任务后，就可以进行任务编排了，代码如下

curl --location --request POST 'http://localhost:8080/api/metadata/workflow' \
--header 'Content-Type: application/json' \
--data-raw '{
    "name": "weather_warning",
    "description": "send weather warning message",
    "version": 4,
    "schemaVersion": 2,
    "ownerEmail": "[email protected]",
    "tasks": [
        {
            "name": "queryWeather",
            "taskReferenceName": "weather",
            "inputParameters": {
                "city": "${workflow.input.city}"
            },
            "type": "SIMPLE"
        },
        {
            "name": "switch_task",
            "taskReferenceName": "is_warning",
            "inputParameters": {
                "temperature": "${weather.output.temperature}"
            },
            "type": "SWITCH",
            "evaluatorType": "javascript",
            "expression": "$.temperature > 37 ? 'Warning' : ''",
            "decisionCases": {
                "Warning": [
                    {
                        "name": "sendMessage",
                        "taskReferenceName": "message",
                        "inputParameters": {
                            "receiver": "${workflow.input.receiver}",
                            "content": "${workflow.input.city}气温为${weather.output.temperature}度,请注意防暑!"
                        },
                        "type": "SIMPLE"
                    }
                ]
            }
        }
    ]
}'

提交成功后，可以在localhost:5000/workflowDef页面看到

workflowDef.png

workflowUi.png

启动工作流

curl --location --request POST 'http://localhost:8080/api/workflow' \
--header 'Content-Type: application/json' \
--data-raw '{
    "name": "weather_warning",
    "version": 4,
    "correlationId": "my_weather_warning_workflows",
    "input": {
        "receiver": "张三",
        "city": "广州"
    }
}'

启动成功后可以在localhost:5000看到在执行的任务，点击进入可以看到执行情况

execute.png

任务实现

前面为了叙事流畅，没有介绍任务的实现。如果没有对应任务实现，上面启动流程后，流程是不会往下执行，它会等待第一个任务的响应。下面用Spring Boot实现 conductor client(worker)

在pom.xml引入

    com.netflix.conductor
    conductor-client-spring
    3.3.6



    com.netflix.conductor
    conductor-common
    3.3.6



    com.netflix.conductor
    conductor-client
    3.3.6

在配置文件application.properties加上配置

conductor.worker.pollingInterval=2
conductor.client.rootURI=http://localhost:8080/api/
conductor.client.threadCount=2

QueryWeatherWorker

import com.netflix.conductor.client.worker.Worker;
import com.netflix.conductor.common.metadata.tasks.Task;
import com.netflix.conductor.common.metadata.tasks.TaskResult;
import org.springframework.stereotype.Component;

import java.util.HashMap;
import java.util.Map;

/**
 * @author zengxc
 */
@Component
public class QueryWeatherWorker implements Worker {
    Map cityTemp = new HashMap<>();
    private final String taskDefName = "queryWeather";

    public QueryWeatherWorker() {
        cityTemp.put("广州", 38);
        cityTemp.put("湖南", 18);
    }

    @Override
    public String getTaskDefName() {
        return taskDefName;
    }

    @Override
    public TaskResult execute(Task task) {
        System.out.printf("Executing %s%n", taskDefName);
        String city = (String) task.getInputData().get("city");
        System.out.println(city + " 气温:" + cityTemp.get(city));

        TaskResult result = new TaskResult(task);
        result.setStatus(TaskResult.Status.COMPLETED);
        //Register the output of the task
        result.getOutputData().put("temperature", cityTemp.get(city));
        result.log(city + " 气温:" + cityTemp.get(city));
        return result;
    }

}

SendMessageWorker

import com.netflix.conductor.client.worker.Worker;
import com.netflix.conductor.common.metadata.tasks.Task;
import com.netflix.conductor.common.metadata.tasks.TaskResult;
import org.springframework.stereotype.Component;

/**
 * @author zengxc
 */
@Component
public class SendMessageWorker implements Worker {

    private final String taskDefName="sendMessage";

    @Override
    public String getTaskDefName() {
        return taskDefName;
    }

    @Override
    public TaskResult execute(Task task) {
        System.out.printf("Executing %s\n", taskDefName);
        System.out.println("接收人:" + task.getInputData().get("receiver")+" "+task.getInputData().get("content"));

        TaskResult result = new TaskResult(task);
        result.setStatus(TaskResult.Status.COMPLETED);
        result.log("接收人:" + task.getInputData().get("receiver")+" "+task.getInputData().get("content"));
        return result;
    }
}