Last value(metric count) of a range vector

[Baitanik Talukder]

I need the last value in a range vector.

Something like 

last_value(failure_count{job="xyz"}[5d])

it is need for the following scenrio:

for last 4 days there was no failure_count metric pushed.. so last value we have 5 days back.

output of 

failure_count{job="xyz"}[5d]

I need the last value (5) from here.

failure_count{endpoint="http",job="xyz"}

8 @1591776884.502 1...@1591776914.502 9 @1591776944.502 5 @1591776974.502

[Julien Pivotto]

On 15 Jun 01:18, Baitanik Talukder wrote:
> I need the last value in a range vector.
> Something like
>

> *last_value*(failure_count{job="xyz"}[5d])

>
> it is need for the following scenrio:
> for last 4 days there was no failure_count metric pushed.. so last value we
> have 5 days back.
> output of
> failure_count{job="xyz"}[5d]
>
> I need the last value (5) from here.
>
> failure_count{endpoint="http",job="xyz"} 8 @1591776884.502
> 1...@1591776914.502
> 9 @1591776944.502
> 5 @1591776974.502

This feature has been requested many times

https://github.com/prometheus/prometheus/pull/1139
https://github.com/prometheus/prometheus/pull/6530
https://github.com/prometheus/prometheus/issues/7179
https://groups.google.com/g/prometheus-developers/c/_h6_QC1zC8U/m/815U0fxQBAAJ

But no consensus has been reached yet to accept that in Prometheus.

In the mean time you can +/- workaround this in recording rules
https://roidelapluie.be./blog/2019/02/21/prometheus-last/

>
> --
> 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 prometheus-use...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/36ba6d12-ec84-43c3-980e-a38b5ce5e873o%40googlegroups.com.


--
Julien Pivotto
@roidelapluie

[Baitanik Talukder]

Okay.. if we try to do the workaround with recording rule.. please tell me how in this case..

just to get the latest know values from this blog (https://roidelapluie.be./blog/2019/02/21/prometheus-last/)

only this one is enough .. right ?

- record: available_sms_last expr: available_sms or available_sms_last

where available_sms is the actual metric

so my case I only have to add :

groups:

- name: test-rules rules: - expr: failure_count or failure_count_last record: failure_count_last

  - alert: failure_count_last_alert

               expr: failure_count_last_alert > 0

[Brian Brazil]

On Mon, 15 Jun 2020 at 09:18, Baitanik Talukder <baitanik...@gmail.com> wrote:

I need the last value in a range vector.

Something like 

last_value(failure_count{job="xyz"}[5d])

it is need for the following scenrio:

 

for last 4 days there was no failure_count metric pushed..

Can you explain more about what you mean about pushing a metric? Anything you pushed to the pushgateway should stay there forever, so it sounds like the issue here could be you using something suboptimally.

Brian

 

so last value we have 5 days back.

output of 

failure_count{job="xyz"}[5d]

I need the last value (5) from here.

failure_count{endpoint="http",job="xyz"}

8 @1591776884.502 1...@1591776914.502 9 @1591776944.502 5 @1591776974.502

--
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 prometheus-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/36ba6d12-ec84-43c3-980e-a38b5ce5e873o%40googlegroups.com.

--

Brian Brazil

www.robustperception.io

[Baitanik Talukder]

that max_over_time or min_over_time can not be applied in my case , because I need recent value..

not the max or min value in that period..

[Baitanik Talukder]

Yeah it is pushed to pushgateway..

so here for last 4 days no metric was pushed to gateway.

Basically I need to check the last 3 known values and based on that raise alert.

most recent count > 0 : raise alert warning

most recent 2 consecutive count > 0 : raise alert severe

most recent 3 consecutive count > 0 : raise alert critical 

How do I do it in prometheus rule ?

On Monday, 15 June 2020 13:48:23 UTC+5:30, Baitanik Talukder wrote:

[Brian Brazil]

On Mon, 15 Jun 2020 at 12:41, Baitanik Talukder <baitanik...@gmail.com> wrote:

Yeah it is pushed to pushgateway..

so here for last 4 days no metric was pushed to gateway.

Basically I need to check the last 3 known values and based on that raise alert.

most recent count > 0 : raise alert warning

most recent 2 consecutive count > 0 : raise alert severe

most recent 3 consecutive count > 0 : raise alert critical 

 

How do I do it in prometheus rule ?

Prometheus is not an event logging system, so that would be quite difficult.

The usual way to do this is to push a gauge when things succeed whose value is the current time - and then alert when that's too long ago by comparing with time().

Brian

 

On Monday, 15 June 2020 13:48:23 UTC+5:30, Baitanik Talukder wrote:

I need the last value in a range vector.

Something like 

last_value(failure_count{job="xyz"}[5d])

it is need for the following scenrio:

for last 4 days there was no failure_count metric pushed.. so last value we have 5 days back.

output of 

failure_count{job="xyz"}[5d]

I need the last value (5) from here.

failure_count{endpoint="http",job="xyz"}

8 @1591776884.502 1...@1591776914.502 9 @1591776944.502 5 @1591776974.502

--
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 prometheus-use...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/ac7783e3-6ae9-4569-a7fd-7b5ee9470663o%40googlegroups.com.

--

Brian Brazil

www.robustperception.io