You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@nifi.apache.org by Pierre Villard <pi...@gmail.com> on 2023/01/02 09:01:33 UTC
Re: Nifi flowfile monitoring
I think it'd be good to take a step back. How can a flow file get stuck in
a relationship? 1 - Is it because you have a self loop failure
relationship? 2 - Is it because you have "dead ends" with a failure
relationship going to a funnel? 3 - or is it because a flow file is failing
and is being rolled back in the input relationship?
With 1, you should leverage the retry feature that has been recently added
With 2, this is a bad practice but you could indeed using reporting tasks
to check flow files in those specific queues
With 3, bulletins would be generated and I recommend having monitoring in
place for the bulletins (with reporting tasks)
Le sam. 31 déc. 2022 à 16:49, Jorge Machado <jo...@me.com> a écrit :
> Activate prometheus exporter and you will have a metric for number of
> objects in the queue
>
> On 31. Dec 2022, at 07:08, nayan sharma <na...@gmail.com> wrote:
>
> Hi Users,
> Is there anyway through which I can monitor or raise alert if any flow
> file got stuck in nifi queue.
>
> For now operation team needs to manually check for these. If you can
> suggest way through which I can achieve this that would be great.
>
> Thanks,
> Nayan
> --
> Thanks & Regards,
> Nayan Sharma
> *+91-8095382952*
>
> <https://www.linkedin.com/in/nayan-sharma>
> <http://stackoverflow.com/users/3687426/nayan-sharma?tab=profile>
>
>
>
Re: Nifi flowfile monitoring
Posted by Guillermo Muñoz <gu...@gmail.com>.
Hi,
We have a seven nodes cluster with 1.16.3. We haven't been able to
reproduce it. We just know that only it happens in the same nodes of the
cluster, but we havent seen any difference between them.
All the processor groups have only one input port and one output port. We
have single batch per node as "Process Group FlowFile Concurrency"
configuration and Batch output as Process Group Outbound Policy.
Thanks. Regards.
El lun, 9 ene 2023 a las 11:46, Pierre Villard (<pi...@gmail.com>)
escribió:
> Hi,
>
> We should figure out why this happens as it should not. Which version are
> you using? Is it something you've been able to reproduce?
> How are flow files getting into the process group, and how are they going
> out? Multiple input ports? Multiple output ports?
> Also you configured "Batch" for the outbound policy but what did you
> configure for the flow file concurrency at process group level?
>
> Pierre
>
>
> Le lun. 9 janv. 2023 à 11:38, Guillermo Muñoz Salgado <mu...@gmail.com>
> a écrit :
>
>> Hi,
>>
>> Sometimes we have flowfiles stuck without reason in queues before out
>> ports in consumer groups with batch output flowfile concurrency configured.
>> We don't know why, and we haven't been able to find out where the problem
>> is. So our operation teams has to check periodically, and change the
>> flowfile concurrency to streaming to unstack these flowfiles.
>>
>> Thanks to all.
>>
>> El lun, 2 ene 2023 a las 10:02, Pierre Villard (<
>> pierre.villard.fr@gmail.com>) escribió:
>>
>>> I think it'd be good to take a step back. How can a flow file get stuck
>>> in a relationship? 1 - Is it because you have a self loop failure
>>> relationship? 2 - Is it because you have "dead ends" with a failure
>>> relationship going to a funnel? 3 - or is it because a flow file is failing
>>> and is being rolled back in the input relationship?
>>> With 1, you should leverage the retry feature that has been recently
>>> added
>>> With 2, this is a bad practice but you could indeed using reporting
>>> tasks to check flow files in those specific queues
>>> With 3, bulletins would be generated and I recommend having monitoring
>>> in place for the bulletins (with reporting tasks)
>>>
>>> Le sam. 31 déc. 2022 à 16:49, Jorge Machado <jo...@me.com> a écrit :
>>>
>>>> Activate prometheus exporter and you will have a metric for number of
>>>> objects in the queue
>>>>
>>>> On 31. Dec 2022, at 07:08, nayan sharma <na...@gmail.com>
>>>> wrote:
>>>>
>>>> Hi Users,
>>>> Is there anyway through which I can monitor or raise alert if any flow
>>>> file got stuck in nifi queue.
>>>>
>>>> For now operation team needs to manually check for these. If you can
>>>> suggest way through which I can achieve this that would be great.
>>>>
>>>> Thanks,
>>>> Nayan
>>>> --
>>>> Thanks & Regards,
>>>> Nayan Sharma
>>>> *+91-8095382952*
>>>>
>>>> <https://www.linkedin.com/in/nayan-sharma>
>>>> <http://stackoverflow.com/users/3687426/nayan-sharma?tab=profile>
>>>>
>>>>
>>>>
>>
>> --
>> Guillermo Muñoz Salgado
>>
>
--
Guillermo Muñoz Salgado
Re: Nifi flowfile monitoring
Posted by Pierre Villard <pi...@gmail.com>.
Hi,
We should figure out why this happens as it should not. Which version are
you using? Is it something you've been able to reproduce?
How are flow files getting into the process group, and how are they going
out? Multiple input ports? Multiple output ports?
Also you configured "Batch" for the outbound policy but what did you
configure for the flow file concurrency at process group level?
Pierre
Le lun. 9 janv. 2023 à 11:38, Guillermo Muñoz Salgado <mu...@gmail.com> a
écrit :
> Hi,
>
> Sometimes we have flowfiles stuck without reason in queues before out
> ports in consumer groups with batch output flowfile concurrency configured.
> We don't know why, and we haven't been able to find out where the problem
> is. So our operation teams has to check periodically, and change the
> flowfile concurrency to streaming to unstack these flowfiles.
>
> Thanks to all.
>
> El lun, 2 ene 2023 a las 10:02, Pierre Villard (<
> pierre.villard.fr@gmail.com>) escribió:
>
>> I think it'd be good to take a step back. How can a flow file get stuck
>> in a relationship? 1 - Is it because you have a self loop failure
>> relationship? 2 - Is it because you have "dead ends" with a failure
>> relationship going to a funnel? 3 - or is it because a flow file is failing
>> and is being rolled back in the input relationship?
>> With 1, you should leverage the retry feature that has been recently added
>> With 2, this is a bad practice but you could indeed using reporting tasks
>> to check flow files in those specific queues
>> With 3, bulletins would be generated and I recommend having monitoring in
>> place for the bulletins (with reporting tasks)
>>
>> Le sam. 31 déc. 2022 à 16:49, Jorge Machado <jo...@me.com> a écrit :
>>
>>> Activate prometheus exporter and you will have a metric for number of
>>> objects in the queue
>>>
>>> On 31. Dec 2022, at 07:08, nayan sharma <na...@gmail.com> wrote:
>>>
>>> Hi Users,
>>> Is there anyway through which I can monitor or raise alert if any flow
>>> file got stuck in nifi queue.
>>>
>>> For now operation team needs to manually check for these. If you can
>>> suggest way through which I can achieve this that would be great.
>>>
>>> Thanks,
>>> Nayan
>>> --
>>> Thanks & Regards,
>>> Nayan Sharma
>>> *+91-8095382952*
>>>
>>> <https://www.linkedin.com/in/nayan-sharma>
>>> <http://stackoverflow.com/users/3687426/nayan-sharma?tab=profile>
>>>
>>>
>>>
>
> --
> Guillermo Muñoz Salgado
>
Re: Nifi flowfile monitoring
Posted by Guillermo Muñoz Salgado <mu...@gmail.com>.
Hi,
Sometimes we have flowfiles stuck without reason in queues before out ports
in consumer groups with batch output flowfile concurrency configured. We
don't know why, and we haven't been able to find out where the problem is.
So our operation teams has to check periodically, and change the flowfile
concurrency to streaming to unstack these flowfiles.
Thanks to all.
El lun, 2 ene 2023 a las 10:02, Pierre Villard (<pi...@gmail.com>)
escribió:
> I think it'd be good to take a step back. How can a flow file get stuck in
> a relationship? 1 - Is it because you have a self loop failure
> relationship? 2 - Is it because you have "dead ends" with a failure
> relationship going to a funnel? 3 - or is it because a flow file is failing
> and is being rolled back in the input relationship?
> With 1, you should leverage the retry feature that has been recently added
> With 2, this is a bad practice but you could indeed using reporting tasks
> to check flow files in those specific queues
> With 3, bulletins would be generated and I recommend having monitoring in
> place for the bulletins (with reporting tasks)
>
> Le sam. 31 déc. 2022 à 16:49, Jorge Machado <jo...@me.com> a écrit :
>
>> Activate prometheus exporter and you will have a metric for number of
>> objects in the queue
>>
>> On 31. Dec 2022, at 07:08, nayan sharma <na...@gmail.com> wrote:
>>
>> Hi Users,
>> Is there anyway through which I can monitor or raise alert if any flow
>> file got stuck in nifi queue.
>>
>> For now operation team needs to manually check for these. If you can
>> suggest way through which I can achieve this that would be great.
>>
>> Thanks,
>> Nayan
>> --
>> Thanks & Regards,
>> Nayan Sharma
>> *+91-8095382952*
>>
>> <https://www.linkedin.com/in/nayan-sharma>
>> <http://stackoverflow.com/users/3687426/nayan-sharma?tab=profile>
>>
>>
>>
--
Guillermo Muñoz Salgado