trigger是Node-RED中的一个常用组件，这篇文章继续结合实例介绍使用trigger进行防止抖动或者消息队列机制的方法。
以容器方式启动Node-RED服务，启动命令如下所示：
启动命令：docker run -it -p 1880:1880 -v $PWD/data:/data -e TZ=Asia/Shanghai --name nodered -d nodered/node-red:1.0.4
限速策略：消息队列 vs 消息抛弃限速必然导致数据无法及时处理完毕，对于这些消息一般有两种做法较为常见：
- 消息队列：按照消息的顺序逐条进行
- 消息抛弃：只处理一定时间段内的第一条，其余的全部抛弃
消息队列是缺省的对应方式，限速之后会对消息队列里的消息逐条处理，很多场景都是可以理解的。而消息抛弃的机制则是抛弃指定时间段内的其余信息，比如同样的操作，连续点击多次，有可能是误操作，实际上一定时间内只需要一此操作就可以了，比如代码提交引起的自动构建，无论是谁的代码提交进行的构建，对这个阶段所有的代码提交者，这次构建的结果都是可以共用的，这种情况下就完全可以使用消息抛弃的机制。比如Jenkins中在很早的版本就是使用类似的方式来控制Rush Hour的性能瓶颈。
消息队列方式 flow设定
使用如下flow来对消息队列方式的限速进行说明
-
输入消息1-3节点：以输入消息1为例，进行说明，为Inject节点的缺省设定
-
limit 1 msg/3s：使用trigger缺省设定，间隔设定为3s
- 结果显示节点：Debug节点的缺省设定
进行部署操作并，手动执行（每个Inject节点快速点击几下）确认结果如下：
和消息队列方式相比，实际上只是多选了一个复选框，详细信息如下：
flow设定
使用如下flow来对消息抛弃方式的限速进行说明
-
输入消息1-3节点：以输入消息1为例，进行说明，为Inject节点的缺省设定
-
结果显示节点：Debug节点的缺省设定
进行部署操作并，手动执行（每个Inject节点快速点击几下）确认结果如下：
[{"id":"891040ad.e0b96","type":"inject","z":"73554227.4474dc","name":"","topic":"","payload":"输入消息2","payloadType":"str","repeat":"","crontab":"","once":false,"onceDelay":0.1,"x":260,"y":160,"wires":[["c6f6eda1.50745"]]},{"id":"c6f6eda1.50745","type":"delay","z":"73554227.4474dc","name":"","pauseType":"rate","timeout":"5","timeoutUnits":"seconds","rate":"1","nbRateUnits":"3","rateUnits":"second","randomFirst":"1","randomLast":"5","randomUnits":"seconds","drop":true,"x":480,"y":160,"wires":[["98da75ae.e12448"]]},{"id":"98da75ae.e12448","type":"debug","z":"73554227.4474dc","name":"结果显示","active":true,"tosidebar":true,"console":false,"tostatus":false,"complete":"payload","targetType":"msg","x":680,"y":160,"wires":[]},{"id":"140a50de.1f611f","type":"inject","z":"73554227.4474dc","name":"","topic":"","payload":"输入消息3","payloadType":"str","repeat":"","crontab":"","once":false,"onceDelay":0.1,"x":260,"y":240,"wires":[["c6f6eda1.50745"]]},{"id":"e7207c13.5a627","type":"inject","z":"73554227.4474dc","name":"","topic":"","payload":"输入消息1","payloadType":"str","repeat":"","crontab":"","once":false,"onceDelay":0.1,"x":260,"y":80,"wires":[["c6f6eda1.50745"]]}]
