testflow | Command line testing tool for Alexa skill code

IntRange.asFlow uses unsafeFlow internally which is defined as:

You are using Lifecycle.State.STARTED state to start observing Flow, the corresponding method of the Activity when emission stops is onStop(). If onStop() method of Activity is called the emission and collecting will stop.

If you want to stop emitting and collection data when onPause method is called, you can use Lifecycle.State.RESUMED state.

When app goes to background onStop() method of Activity is called and when using Lifecycle.State.STARTED state to observe the Flow you should see the emission and collecting stop.

This is a know issue in Mule 3.9.0 to Mule 3.9.4. The solution is to upgrade to Mule 3.9.5.

Source: https://help.mulesoft.com/s/article/Store-procedure-calls-fail-in-mule-3-x-for-snowflake-database

If stop, but not destroy, everything in the process is going to be handled properly. Only new data is going to be emitted from the source or message channels. This is also natural graceful shutdown behavior of an ApplicationContext when you close it: it stops beans first letting them to finish whatever is in progress an only after that it destroys them.

Therefore so far you should be OK with your intentions. Share with us, please, your experience if it otherwise - and we will look into that ASAP with an appropriate fix if needed. The point is that it really was a goal during the stop phase to not emit new data, but let existing processed to finish gracefully. See more info in docs: https://docs.spring.io/spring-integration/docs/current/reference/html/system-management.html#jmx-shutdown

According your current configuration the testChannel is a DrectChannel, so whatever you send to it is going to be processed on a thread your send from. Therefore the Thread.currentThread() is enough for your to determine it.

For more general solution consider to have a MessagePublishingErrorHandler as a bean with the ChannelUtils.MESSAGE_PUBLISHING_ERROR_HANDLER_BEAN_NAME to override a default one. This MessagePublishingErrorHandler can be supplied with a custom ErrorMessageStrategy. There, when you create an ErrorMessage, you can add a custom header with the same Thread.currentThread() info to carry onto that error channel processing even if it is done in a separate thread.

You also could just throw an exception with that info, too, instead!

See gateway(IntegrationFlow) method of flow definition :

After having a similar problem using only one level of scatter-gather, I realized it was the log message that was blocking the output from being returned to the parent flow. Replace .log() with .logAndReply() or .log().bridge() and everything should work again.

Like this:

If the execution of the Flow throws an Exception, it will cancel and complete the Flow during collection. The collect() function call will throw the Exception if the Flow.catch operator was not used.

If you emit an Exception like in your example, it's just another object in the Flow. Since you have not specified the Flow's type, it's implicitly choosing a type that's common between String and Exception. I think you have a Flow since that's a common supertype of both. If you had specified Flow, it would not allow you to emit an Exception.

Bug filed, marking this question as answered. Thanks to everyone for the quick responses.