Hi, Julius!
Thanks for your time and I'm sorry, it's my misunderstanding nature of
"promhttp_" metrics.
Currently reread documentation and now everything is clear.
среда, 17 августа 2022 г. в 14:29:38 UTC+3, Oleksandr Matsak:
> Look please at the code below
>
> metricsMux := http.NewServeMux()
> // here I want only to expose metrics outside, without any tracking
> // just gather them and write response
> metricsMux.Handle("/metrics", promhttp.Handler()) // <------
> metricsSrv := &http.Server{Handler: metricsMux}
>
> appMux := http.NewServeMux()
> // here I want to add prometheus middleware with "promhttp_"-prefix
> tracking
> // only tracking, without any output written
>
> appSrv := &http.Server{Handler: appMux}
>
>
> If simply add promhttp.Handler() handler to metrics server(as in my
> example), "promhttp_"-prefix metrics will contains info about requests to
> metrics server and it's useless information. I want them to contain metrics
> for my application.
>
> Short: I'm trying to follow separation of concerns while tracking and
> exposing metrics.
>
> среда, 17 августа 2022 г. в 13:38:17 UTC+3, [email protected]:
>
>> Hi Oleksandr,
>>
>> I'm a bit confused. The "promhttp_"-prefixed metrics are about the scrape
>> requests themselves and they are still tracked fine when tying the
>> promhttp.Handler() to another mux and server (because the scrape requests
>> are now happening through that server, and the handler is also doing its
>> tracking there now). Or are you talking about HTTP metrics for the rest of
>> your application's HTTP handlers (not /metrics) that is running on the
>> other server?
>>
>> On Wed, Aug 17, 2022 at 12:00 PM Oleksandr Matsak <[email protected]>
>> wrote:
>>
>>> Hi, Julius!
>>> Thanks for quick reply.
>>>
>>> Yes, you are right, if I will use separate http.ServeMux on separate
>>> server I will loose "promhttp_*" metrics for my application(because they
>>> will be counted for separate server).
>>>
>>>
>>> I have found the solution, but don't like it.
>>>
>>>
>>> вторник, 16 августа 2022 г. в 21:05:14 UTC+3, [email protected]:
>>>
>>>> Hi Alex,
>>>>
>>>> both approaches are generally fine, based on your needs. With "losing
>>>> instrumentation metrics", do you mean the "promhttp_*" metrics about the
>>>> scrape itself? Those should still be present when just using
>>>> "promhttp.Handler()" on a different http.ServeMux and starting a second
>>>> server for that ServeMux. Could you share the relevant lines of your code
>>>> and/or elaborate?
>>>>
>>>> Regards,
>>>> Julius
>>>>
>>>> On Mon, Aug 15, 2022 at 11:34 PM Oleksandr Matsak <
>>>> [email protected]> wrote:
>>>>
>>>>> Hi!
>>>>>
>>>>> I have REST API application and looking for solution for exposing
>>>>> metrics endpoint for internal only access.
>>>>>
>>>>> I can see two options:
>>>>>
>>>>> 1. use same *http.Server* as application and protect endpoint with
>>>>> supported auth flow (basic, token, etc)
>>>>> 2. use different *http.Server* on internal port without any
>>>>> authentication
>>>>>
>>>>> First has drawback of hardcoding auth credentials on both
>>>>> infrastructure and application sides, I want to avoid it.
>>>>>
>>>>> While considering second, found that if just put *promhttp.Handler*
>>>>> to separate server I will loose instrumentation metrics.
>>>>>
>>>>> So how do I see it:
>>>>>
>>>>> - put instrumentation metrics middleware to application server
>>>>> - run another server on internal only port with metrics output
>>>>> handler
>>>>>
>>>>> Any advise ?
>>>>>
>>>>> Thanks,
>>>>> Alex
>>>>>
>>>>> --
>>>>> You received this message because you are subscribed to the Google
>>>>> Groups "Prometheus Users" group.
>>>>> To unsubscribe from this group and stop receiving emails from it, send
>>>>> an email to [email protected].
>>>>> To view this discussion on the web visit
>>>>> https://groups.google.com/d/msgid/prometheus-users/54038168-acda-40f2-97f7-7fc1212e926cn%40googlegroups.com
>>>>>
>>>>> <https://groups.google.com/d/msgid/prometheus-users/54038168-acda-40f2-97f7-7fc1212e926cn%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>>> .
>>>>>
>>>>
>>>>
>>>> --
>>>> Julius Volz
>>>> PromLabs - promlabs.com
>>>>
>>> --
>>> You received this message because you are subscribed to the Google
>>> Groups "Prometheus Users" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to [email protected].
>>>
>> To view this discussion on the web visit
>>> https://groups.google.com/d/msgid/prometheus-users/1766f79b-ab33-401c-8e0d-f2e7e935ff35n%40googlegroups.com
>>>
>>> <https://groups.google.com/d/msgid/prometheus-users/1766f79b-ab33-401c-8e0d-f2e7e935ff35n%40googlegroups.com?utm_medium=email&utm_source=footer>
>>> .
>>>
>>
>>
>> --
>> Julius Volz
>> PromLabs - promlabs.com
>>
>
--
You received this message because you are subscribed to the Google Groups
"Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/prometheus-users/c0e309f6-3d87-40bd-bace-9f7c4712658en%40googlegroups.com.
