Question

我已经设置了2个节点工作程序和1个节点管理程序siddhi集群。以下是我尝试推送到经理的查询。当查询中没有任何联接时，一切似乎都正常运行，但是，如下所述，如果联接发生，查询将部署在工作程序节点中，但事件似乎无法得到满足。

_layout.cshtml

事件正在传递

@app:name("rule_1")
@source(
  type="kafka",
  topic.list="test-input-topic",
  group.id="test-group",
  threading.option="single.thread",
  bootstrap.servers="localhost:9092",
  @Map(type="json"))
 define stream TempStream (deviceID string,roomNo string,temp int );

 @sink(
  type="kafka",
  topic="test-output-topic",
  bootstrap.servers="localhost:9092",
  @Map(type="json"))
 define stream OutStream (message string, message1 string, message2 double);

 @info(name = "query1")
 @dist(execGroup="group1")
 from TempStream[deviceID=="rule_1" and temp>10]#window.time(5 sec)
 select avg(temp) as avgTemp, roomNo, deviceID
 insert all events into AvgTempStream1;

 @info(name = "query2")
 @dist(execGroup="group2")
 from TempStream[deviceID=="rule_1" and temp<10]#window.time(5 sec)
 select avg(temp) as avgTemp, roomNo, deviceID
 insert all events into AvgTempStream2;

 @info(name = "query3")
 @dist(execGroup="group3")
 from AvgTempStream1#window.length(1) as stream1 join AvgTempStream2#window.length(1) as stream2
 select stream1.deviceID as message,stream1.roomNo as message1, stream1.avgTemp as message2
 having stream1.avgTemp>stream2.avgTemp
 insert into outputStream;

 @info(name = "query4")
 @dist(execGroup="group4")
 from AvgTempStream1[deviceID=="rule_1"]
 select deviceID as message, roomNo as message1, avgTemp as message2
 insert into OutStream;

Answer 1

根据上述给定的Siddhi应用程序，您将需要发送两个输入事件以使其输出。一个事件的温度> 10，其他温度<10。例如：

{"event":{"deviceID":"rule_1","roomNo":"123","temp":12}}

{"event":{"deviceID":"rule_1","roomNo":"123","temp":8}}

这将确保发生联接并发出事件。出于故障排除的目的，您可以使用Kafka使用者订阅不确定的Kafka主题。中间主题的名称将遵循SiddhiAppName.StreamName的格式。例如：rule_1.AvgTempStream1

希望这会有所帮助！

谢谢，提山

与Join操作一起对siddhi查询进行完全分布式部署的问题

1 个答案: