org.mule.runtime.core.internal.event.EventContextDeepNestingException

SYMPTOM

You are using the Mule Runtime 4.x.x and you seeing the following exception in the log files. 
 

Root Exception stack trace:
org.mule.runtime.core.internal.event.EventContextDeepNestingException: Too many child contexts nested.

    at org.mule.runtime.core.internal.event.AbstractEventContext.addChildContext(AbstractEventContext.java:117)
    at org.mule.runtime.core.internal.event.DefaultEventContext.child(DefaultEventContext.java:75)
    at org.mule.runtime.core.internal.event.DefaultEventContext.child(DefaultEventContext.java:56)
    at org.mule.runtime.core.privileged.processor.MessageProcessors.newChildContext(MessageProcessors.java:238)
    at org.mule.runtime.core.privileged.processor.MessageProcessors.processWithChildContext(MessageProcessors.java:226)
    at org.mule.runtime.config.internal.factories.FlowRefFactoryBean$FlowRefMessageProcessor.lambda$apply$0(FlowRefFactoryBean.java:218)
    at reactor.core.publisher.FluxFlatMap$FlatMapMain.onNext(FluxFlatMap.java:347)
    at org.mule.runtime.core.privileged.processor.chain.AbstractMessageProcessorChain$1.onNext(AbstractMessageProcessorChain.java:255)
    ... followed by a long stacktrace.

CAUSE

There is a (currenty undocumented) setting org.mule.runtime.core.privileged.event.BaseEventContext.maxDepth that has introduced a limitation on the number of invocations that you can have to a (sub)flow from a parent flow. This was done to prevent unnecessary stack overflows being generated when flows are called to frequently and too many contexts being created. By default, the setting org.mule.runtime.core.privileged.event.BaseEventContext.maxDepth is set to be 25, but can be increased.

SOLUTION

1) Check the flows to ensure that the references to other flows is done without unnecessary recursion.

2) You can design alternative to a recursive flow call, you can have the following flow logic:
a) a VM listener flow that listens to a VM queue. You can initially populate this VM queue to trigger the process. At the end of this flow, you can decide whether to push a payload back into the VM queue (thus triggering the flow again), or breaking out of it altogether. This will avoid the long stack traces causing the nested context error.

We have attached a sample app in Mule 4 that will trigger counting from 1 - 10. Each time, it will aggregate the variable and pushes to the VM queue. If it goes over 10, it will stop the flow.
Run the following command to trigger the sample app running inside Studio:

curl localhost:8081

 Please note, the connector versions will need to be reviewed and updated as needed in .jar file

Attachments

vm-queue-recursive-alternative.jar