[英]How to Interrupt or Cancel a Spring Integration Java DSL flow?
Contemplating implementation of a feature wherein a JdbcRepositoryHandler
(implemnting MessageHandler
) might listen for an outside event (eg, CancelRunEvent
). 考虑某个功能的实现,其中
JdbcRepositoryHandler
(实现MessageHandler
)可能会监听外部事件(例如CancelRunEvent
)。
I think I'd use Spring' ApplicationEvent
support to publish an event via a REST controller endpoint. 我想我会使用Spring的
ApplicationEvent
支持来通过REST控制器端点发布事件。 And I would guess I'd have the aforementioned handler implement ApplicationListener
to listen for a particular event? 我猜想我会使用前面提到的处理程序实现
ApplicationListener
来监听特定事件吗?
The question is: if the handler is saturated with messages it needs to process, how would I signal termination of all subsequent messages that may have emanated upstream eg, from a FileSplitter
? 问题是: 如果处理程序中充满了需要处理的消息,我该如何终止可能从上游发出的所有后续消息(例如,来自
FileSplitter
信号)的终止 ?
While I could easily construct a condition to be checked before calling a method responsible eg, for a persistence operation (based on some state received from the CancelRunEvent
), how could I interrupt the flow entirely ? 虽然我可以在调用负责持久性操作的方法(基于从
CancelRunEvent
接收到的某些状态)之前轻松构造要检查的条件,但如何完全中断流程 ?
For illustrative purposes, imagine a flow like: 出于说明目的,请设想如下流程:
@Bean
protected IntegrationFlow s3ChannelFlow() {
// @see http://docs.spring.io/spring-integration/reference/html/files.html#file-reading
// @formatter:off
return IntegrationFlows
.from(s3Channel())
.enrichHeaders(h -> h.headerFunction(RunStats.FILE_TOKEN, f -> UUID.randomUUID().toString()))
.channel(runStatsChannel())
.transform(new FileToInputStreamTransformer())
.split(new FileSplitter())
.transform(new JsonToObjectViaTypeHeaderTransformer(new Jackson2JsonObjectMapper(objectMapper), typeSupport))
.publishSubscribeChannel(p -> p.subscribe(persistenceSubFlow()))
.get();
// @formatter:on
}
@Bean
protected IntegrationFlow persistenceSubFlow() {
// @formatter:off
return f -> f
// @see http://docs.spring.io/spring-integration/reference/html/messaging-routing-chapter.html#agg-and-group-to
.aggregate(a -> a
.correlationStrategy(new HeaderAttributeCorrelationStrategy(RunStats.FILE_TOKEN))
.releaseStrategy(new MessageCountReleaseStrategy(persistenceBatchSize))
.sendPartialResultOnExpiry(true)
.expireGroupsUponCompletion(true)
.groupTimeoutExpression(persistenceBatchReleaseTimeoutMillis)
)
.handle(new JdbcRepositoryHandler(typeSupport, metricRegistry, runStatsRepository));
// @formatter:on
}
It's not entirely clear what you mean or why you would need the JdbcRepositoryHandler
to manage this rather than some other ApplicationListener
. 目前尚不清楚您的意思是什么,或者为什么需要
JdbcRepositoryHandler
来管理它而不是其他ApplicationListener
。
Your flow is running on some thread upstream of s3Channel()
. 您的流在
s3Channel()
上游的某个线程上运行。 Depending on what that is, you could stop()
the message source and no new messages will be emitted after the current one (or ones if it's multi-threaded). 根据具体情况,您可以
stop()
消息源,并且在当前消息(或多线程消息)之后不会发出新消息。
However, you may (likely will) end up with a partial aggregation sitting in memory until the group times out. 但是,您可能(可能会)最终在内存中停留部分聚合,直到组超时。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.