Hello fellow NiFi users,

we have some fairly large NiFi flows and would like to monitor whether
there are FlowFiles stucked, e.g. because of failing processors or being
stuck in a loop.

Is there a way to find any FlowFiles being "stuck" in the Flow for longer
than a time period of X, e.g. an hour at any part of the flow? I'm aware of
the concept of "Lineage duration" but could not find a REST endpoint or
something similar that would allow this to be checked.

We're not interested in monitoring overall processing times but rather want
to ensure that all FlowFiles are processed successfully within a certain
time frame.
*Is there a way to get the lineage duration of all FlowFiles currently in
the whole flow?*

We thought about crawling the REST api for every connection and listing
their queues but this seems rather inconvenient. Also we would not be able
to track down files currently being processed by a processor.
Is there an easier way?

Monitoring the provenance events does not seem to be an option as this
would miss FlowFiles being stuck in a queue for a long period of time as
this does not create a provenance event.
Also we would not be able to a) pin-point where the FlowFile is at and b)
more importantly determine whether the FlowFile is still in the flow or
processed successfully yet.

Is there some alternative?

Has anyone an idea whether this would be hard to bring to NiFi itself?
I'm not familiar enough with the source code of NiFi to assess the
difficulty.

Any hints are highly appreciated!
Thank you. :)

Reply via email to