Digging a little deeper I see that a Funnel gets its active thread count from LifecycleState which is initialized with an activeThreadCount of 0. I do not believe Funnels ever update this. So, the short answer is that they do not count towards the active (and therefore available) thread count.
The question arose due to backpressure building on a connection leading into a funnel. I was curious how this could ever happen. (The output connection of the funnel has a backpressure setting of 0 object/0 bytes.) On Fri, Feb 8, 2019 at 10:40 AM Joe Witt <joe.w...@gmail.com> wrote: > Mark > > I believe it does pull from the same thread pool but not sure if it is > 'counted'. > > Alternatively it could be implemented such that whenever data is written to > a queue we check if the target of that queue is a funnel and if so then > immediately do the right 'next' thing until the next target isn't a > funnel. Could do the same with local ports.. > > Thanks > > On Fri, Feb 8, 2019 at 10:23 AM Mark Bean <mark.o.b...@gmail.com> wrote: > > > It seems that funnels should not require a thread; they do not actually > > process data. However, looking into the code for both 1.8.0 [1] and > master > > [2], it seems that a thread is allocated to each funnel when gathering > the > > status of a Process Group. Does this count against threads available to > be > > allocated to processors? > > > > [1] > > > > > https://github.com/apache/nifi/blob/98aabf2c50f857efc72fd6f2bfdd9965b97fa195/nifi-nar-bundles/nifi-framework-bundle/nifi-framework/nifi-framework-core/src/main/java/org/apache/nifi/controller/FlowController.java#L3326 > > [2] > > > > > https://github.com/apache/nifi/blob/e4be06c7c8a4801e375344954e442bb15aa3f636/nifi-nar-bundles/nifi-framework-bundle/nifi-framework/nifi-framework-core/src/main/java/org/apache/nifi/controller/FlowController.java#L1936 > > > > Thanks, > > Mark > > >
