Hi,

Is it possible to have a common relabel_config configuration (may be under
global configuration) for all jobs instead of duplicating the same
relabel_config under each job?

Thanks,
Anoop


On Thu, Mar 3, 2022 at 2:58 PM Anoop Mohan <[email protected]> wrote:

> Thanks Brian for a good explanation. Will go with target relabelling.
>
> Thanks,
> Anoop
>
>
> On Wed, Mar 2, 2022 at 9:03 PM Brian Candler <[email protected]> wrote:
>
>> Target relabelling occurs *before* the scrape takes place; metrics
>> relabelling occurs *after* the scrape has taken place.
>>
>> If you're just adding fixed labels to every metric, then either way will
>> work.  However I'd expect target relabelling to be more efficient, since
>> it's only run once for each target, whereas metric relabelling is run once
>> for every individual timeseries returned in the scrape.
>>
>> Otherwise:
>> * If you want to set any labels which affect how the scrape is done - for
>> example changing the metrics_path or adding query parameters when
>> communicating with the exporter, or changing the address or port scraped -
>> then this has to be done in target relabelling.
>> * If you want to label some returned timeseries differently to others, or
>> selectively drop timeseries, then this has to be done in metric relabelling.
>>
>>
>> On Wednesday, 2 March 2022 at 15:01:57 UTC Anoop wrote:
>>
>>> Thank you so much Brian. The "*relabel_configs*" worked for me.
>>>
>>> I tried "*metric_relabel_configs*" also with the below configuration
>>> and this is also adding the new label with all metrics. Not sure if this is
>>> the correct method though :)
>>>
>>> metric_relabel_configs:
>>>   - source_labels: [__name__]
>>>     target_label: foo
>>>     replacement: bar
>>>
>>> I am going to use "*relabel_configs*" anyway. However, I'm just curious
>>> whether we can use the "*metric_relabel_configs*" method also? Will it
>>> make any issue?
>>>
>>> Thanks,
>>>
>>>
>>>
>>>
>>>
>>>
>>> Thanks & Regards,
>>>
>>> Anoop Mohan
>>>
>>> Mob# +91-7293009486 <+91%2072930%2009486>
>>> E-mail : [email protected]
>>>
>>>
>>> On Wed, Mar 2, 2022 at 7:38 PM Brian Candler <[email protected]> wrote:
>>>
>>>> If you're using kubernetes_sd_configs
>>>> <https://prometheus.io/docs/prometheus/latest/configuration/configuration/#kubernetes_sd_config>
>>>> then I don't think you can do it this way.
>>>>
>>>> Instead, you'll need to use target relabelling:
>>>>
>>>>   - job_name: foo
>>>>     kubernetes_sd_configs:
>>>>       ...
>>>> *    relabel_configs:*
>>>>
>>>> *      - target_label: foo        replacement: bar*
>>>>
>>>> This should add {foo="bar"} to every target, and hence to every scraped
>>>> metric.
>>>>
>>>> Note that the SD mechanism sets lots of __meta_XXX labels on each
>>>> target, so there's lots of opportunity here to use these values and copy
>>>> them into "real" labels as required.  (Labels which start with
>>>> double-underscore are removed after relabelling).
>>>>
>>>> On Wednesday, 2 March 2022 at 13:50:42 UTC Anoop wrote:
>>>>
>>>>> Thanks Brian for your suggestion.
>>>>>
>>>>> I am using "kubernetes_sd_configs" to discover the targets
>>>>> dynamically. So I just tried the below configuration, since I don't want 
>>>>> to
>>>>> specify any specific targets under "static_configs". But, this is not
>>>>> working as expected as I can't see the custom label in the query results.
>>>>>
>>>>> static_configs:
>>>>>   - targets:
>>>>>     - .*
>>>>>     labels:
>>>>>       foo: bar
>>>>>
>>>>> Not sure if this implementation is correct or not. Please share your
>>>>> suggestions.
>>>>>
>>>>> Thanks
>>>>>
>>>> --
>>>>
>>> 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/e6cc71e1-dbd0-477a-8d57-392a27d38bb8n%40googlegroups.com
>>>> <https://groups.google.com/d/msgid/prometheus-users/e6cc71e1-dbd0-477a-8d57-392a27d38bb8n%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>> .
>>>>
>>> --
>> 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/09f08de8-ff05-4d78-bf91-ef131d3e8854n%40googlegroups.com
>> <https://groups.google.com/d/msgid/prometheus-users/09f08de8-ff05-4d78-bf91-ef131d3e8854n%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
>

-- 
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/CAEQqSpf4%3DkgY3SJdgbKLJ_KY1ADbPrJPp8YfkrRST5pajM9FLA%40mail.gmail.com.

Reply via email to