定期更新的 Flink 源碼

Question

我正在嘗試為長時間運行的 flink 作業實現外部配置。 我的想法是創建自定義源，定期（每 5 分鍾）通過 http 從外部服務輪詢 JSON 編碼的配置。

如何創建每 N 分鍾執行一次操作的源？ 如何將此配置重新廣播給所有執行程序？

Answer 1

首先，您需要創建一個事件類，它將定義您的事件流具有的所有屬性，然后創建所有 getter、setter 和其他方法。 這個類的一個例子是

 public class qrsIntervalStreamEvent {

    public Integer Sensor_id;
    public long time;
    public Integer qrsInterval;


    public qrsIntervalStreamEvent(Integer sensor_id, long time, Integer qrsInterval) {
        Sensor_id = sensor_id;
        this.time = time;
        this.qrsInterval = qrsInterval;
    }


    public Integer getSensor_id() {
        return Sensor_id;
    }

    public void setSensor_id(Integer sensor_id) {
        Sensor_id = sensor_id;
    }

    public long getTime() {
        return time;
    }

    public void setTime(long time) {
        this.time = time;
    }

    public Integer getQrsInterval() {
        return qrsInterval;
    }

    public void setQrsInterval(Integer qrsInterval) {
        this.qrsInterval = qrsInterval;
    }

    @Override
    public boolean equals(Object o) {
        if (this == o) return true;
        if (!(o instanceof qrsIntervalStreamEvent)) return false;

        qrsIntervalStreamEvent that = (qrsIntervalStreamEvent) o;

        if (getTime() != that.getTime()) return false;
        if (getSensor_id() != null ? !getSensor_id().equals(that.getSensor_id()) : that.getSensor_id() != null)
            return false;
        return getQrsInterval() != null ? getQrsInterval().equals(that.getQrsInterval()) : that.getQrsInterval() == null;
    }

    @Override
    public int hashCode() {
        int result = getSensor_id() != null ? getSensor_id().hashCode() : 0;
        result = 31 * result + (int) (getTime() ^ (getTime() >>> 32));
        result = 31 * result + (getQrsInterval() != null ? getQrsInterval().hashCode() : 0);
        return result;
    }


    @Override
    public String toString() {
        return "StreamEvent{" +
                "Sensor_id=" + Sensor_id +
                ", time=" + time +
                ", qrsInterval=" + qrsInterval +
                '}';
    }


} //class

現在假設您想在 x 個事件/ 5 秒發送這些事件，然后您可以編寫這樣的代碼

public class Qrs_interval_Gen extends RichParallelSourceFunction<qrsIntervalStreamEvent> {
@Override
public void run(SourceContext<qrsIntervalStreamEvent> sourceContext) throws Exception {


    int qrsInterval;
    int Sensor_id;
    long currentTime;
    Random random = new Random();

    Integer InputRate = 10;

    Integer Sleeptime = 1000 * 5 / InputRate  ;


    for(int i = 0 ; i <= 100000 ; i++){


        // int randomNum = rand.nextInt((max - min) + 1) + min;
        Sensor_id =  1;

        qrsInterval =  10 + random.nextInt((20-10)+ 1);
       // currentTime = System.currentTimeMillis();
        currentTime = i;

        //System.out.println("qrsInterval = " + qrsInterval + ", Sensor_id = "+ Sensor_id );
        try {
            Thread.sleep(Sleeptime);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }


        qrsIntervalStreamEvent stream = new qrsIntervalStreamEvent(Sensor_id,currentTime,qrsInterval);

            sourceContext.collect(stream);

    } // for loop


}

    @Override
    public void cancel() {

    }
}

這里整個邏輯是由

如果您想發送 x 個事件/秒，那么您的睡眠時間將與此相反。 例如發送 10 個事件/秒

睡眠時間 = 1000 / 10 = 100 毫秒

同樣，對於發送 10 個事件/5 秒，睡眠時間將為

睡眠時間 = 1000 * 5 / 10 = 500 毫秒

希望能幫到你，有什么問題可以私信我