Hello,

Thanks for the question.  The api documentation [1] does provide info about
the endpoint and its parameters, but it doesn't provide the specifics on
exactly what is supported for that parameter. In reviewing the code however
I can see that information (see FlowResource[2]).  I will create a JIRA so
that it can be documented in full for the api.

Best Regards,

-yolanda

[1] NiFi Rest API documentation (search for Flow, getMetrics method) -
https://nifi.apache.org/docs/nifi-docs/rest-api/index.html
[2] FlowResource Implementation -
https://github.com/apache/nifi/blob/rel/nifi-1.23.0/nifi-nar-bundles/nifi-framework-bundle/nifi-framework/nifi-web/nifi-web-api/src/main/java/org/apache/nifi/web/api/FlowResource.java#L418

On Wed, Aug 16, 2023 at 8:33 AM <[email protected]> wrote:

> Hello Yolanda
>
> Where is the document about accepted parameter to set after
> “includedRegistries”
>
> Thanks for help
>
>
> Envoyé à partir de l'app mail mobile
>
> sur 14/08/2023 le 17:38, Yolanda Davis écrivit:
>
> De: "Yolanda Davis" <[email protected]>
> Date: 14 août 2023
> à: [email protected]
> Cc:
> Objet: Re: Node/processor/queue status in grafana
> Hi Aaron,
>
> Great to hear it's something you can use!  You are correct in that it
> should expose everything that the PrometheusReportingTask does.   It also
> has the ability to limit the metrics exposed on the endpoint, however it is
> different from how the reporting task can limit metrics.  So for example
> let's say you only want  to make JVM or NIFi specific metrics available,
> you can pass that as a parameter value to only include the registries for
> those metrics (e.g.
> https://localhost:8443/nifi-api/flow/metrics/prometheus?includedRegistries=JVM).
> There are other filtering mechanisms, but you can also just limit what
> Prometheus scrapes using its own scrape configurations.
>
> One thing also to be mindful of is the amount of metrics you are
> collecting and your storage constraints in Prometheus.  Depending on the
> size of your flow there could be lots of metrics produced for components.
> The amount of space needed for these metrics are driven not just by the
> name and value pair but also the labels associated with it; so you may want
> to consider how to configure your scrape to grab only what is necessary, at
> intervals that best matches your query needs.
>
> Again, I hope this is helpful!
>
> -yolanda
>
> On Mon, Aug 14, 2023 at 1:09 AM Aaron Rich <[email protected]> wrote:
>
>> Hi Yolanda,
>>
>> That is SUPER helpful to know. This looks like a great starting point.
>>
>> It looks like this is equivalent to reporting task for all components. Is
>> that correct? I'll have to see if I can get Prometheus to scrape it with a
>> secure endpoint.
>>
>> Thanks for the starting pointer!
>>
>> -Aaron
>>
>> On Sun, Aug 13, 2023 at 7:26 PM Yolanda Davis <[email protected]>
>> wrote:
>>
>>> Hi Aaron,
>>>
>>> Just wanted to provide some additional info that hopefully will be
>>> useful.  NiFi also has a dedicated prometheus endpoint that can be scraped
>>> directly, without needing to set up a specific task.  That endpoint can be
>>> found (in a local host example) under
>>> https://localhost:8443/nifi-api/flow/metrics/prometheus.  From here you
>>> can see the metrics NiFi makes available and prometheus can be configured
>>> to scrape depending on the security settings you have in place. Again hope
>>> this helps, and good luck!
>>>
>>> -yolanda
>>>
>>> On Sun, Aug 13, 2023 at 8:03 PM Joe Witt <[email protected]> wrote:
>>>
>>>> This list is perfectly fine.
>>>>
>>>> All of our metrics are available via push, pull, and prometheus so it
>>>> can def be done.   It would be great to see what you end up with.
>>>>
>>>> Thanks
>>>>
>>>> On Sun, Aug 13, 2023 at 3:56 PM Aaron Rich <[email protected]>
>>>> wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> Not sure if this belongs on the dev or user mailing list but figured
>>>>> would start with user.
>>>>>
>>>>> I wanted to see if there is a way to create the same graphs available
>>>>> in the status displays using grafana? I'm assuming I would need to use a
>>>>> prometheus reporting task but wanted to know if all the metrics are
>>>>> reported to generate same graphs.
>>>>>
>>>>> I want to be able to graph nifi performance metrics at the same time
>>>>> as my kubernetes cluster metrics so I can understand how they relate and
>>>>> what resources are impacting any slow down in performance of NiFi.
>>>>>
>>>>> If anyone has already attempted this, wouls love to get any pointers
>>>>> for implementing.
>>>>>
>>>>> Thanks.
>>>>>
>>>>> Aaron
>>>>>
>>>>
>>>
>>> --
>>> --
>>> [email protected]
>>> @YolandaMDavis
>>>
>>>
>
> --
> --
> [email protected]
> @YolandaMDavis
>
>

-- 
--
[email protected]
@YolandaMDavis

Reply via email to