Thanks for your observations.
I tried the approach you suggested but I am not getting the expected result
which is an incremental counter. I see the results incrementing when is
less than 98 but I also see decreasing result sometimes when is not less
than 98. It does not make any sense to me.
So I tried a work around with a recording rule to register the times when
the result is less than 98. The recording rule comes from this query:
record: service:http_service:availability:1m
expr:
sum by(method) (100*sum by (method)
(rate(demo_api_request_duration_seconds_count{status="200"}[1m])) / sum by
(method) (rate(demo_api_request_duration_seconds_count[1m])) < bool 98
It works fine. Then , I just do:
sum_over_time(service:agent_dispatcher_api:availability:1m[1d:1m])
I solved my problem from this perspective. But I am still not able to
figure out why I am getting decreasing results from the query in question.
Thanks
Lp
On Friday, April 23, 2021 at 12:00:34 PM UTC-4 [email protected] wrote:
> Hi,
>
> There are at least a couple of things off in that query:
>
> * http_service_api looks like it's probably a counter metric (although it
> doesn't have a "_total" suffix), so you need to apply rate() to it first,
> before applying a sum(). Otherwise you are looking at the absolute value of
> the counter, which has accumulated over a long time, which doesn't tell you
> much.
> * Due to your "by(method)" your output labels are just "method" on both
> sides of the binop, so you don't need the "ignoring()" and also not the
> "group_left()".
>
> Maybe you mean something closer to this?
> https://demo.promlens.com/?l=14PEjXeJocK (changed the metric and label
> names to ones that have data in the demo setup)
>
> sum_over_time(sum by(method) (100*sum by (method)
> (rate(demo_api_request_duration_seconds_count{status="200"}[1m])) / sum by
> (method) (rate(demo_api_request_duration_seconds_count[1m])) < bool
> 98)[1d:1m])
>
> I guess in the end it should give you something close to "how many minutes
> per method over the last day did each method have a percentage of <98 of
> statusCode="0"?
>
> Regards,
> Julius
>
> On Fri, Apr 23, 2021 at 5:17 PM Luciano Polo <[email protected]>
> wrote:
>
>> HI,
>>
>> I am not able to make this query to work precise enough. Hopefully,
>> someone can help me out.
>>
>> This Query works fine. It returns 1 if the result is less than 98
>> 100*sum(http_service_api{apiCode="0"}) by (method) / ignoring (apiCode)
>> group_left() sum(http_service_api{} ) by (method) < bool 98
>>
>> This is the part I am having issues with. I need to add all ones within a
>> period of time where the result was lower than 98.
>> So the query is:
>>
>> sum_over_time(sum(100*sum(http_service_api{apiCode="0"}) by (method) /
>> ignoring (apiCode) group_left() sum(http_service_api{} ) by (method) <
>> bool 98) by (method)[1d:1m])
>>
>> I should expect the number to increase every time the result is less than
>> 98, I this but I also see it decreasing. I have not been able to figure
>> out why?
>>
>> Any idea is appreciated.
>>
>> Thanks in advanced
>>
>> --
>> 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/9decf191-cf6d-4194-a092-d714e2cfdab5n%40googlegroups.com
>>
>> <https://groups.google.com/d/msgid/prometheus-users/9decf191-cf6d-4194-a092-d714e2cfdab5n%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/41c8a3d3-fd84-4275-aad4-d87a2cd6af26n%40googlegroups.com.
