> Try sum() instead of max_over_time().
Do you mean sum_over_time() ? But it will amount to the same thing, surely?
On Tuesday, 16 August 2022 at 15:57:44 UTC+1 Brian Candler wrote:
> What you are saying doesn't make sense, so you need to provide some
> evidence: actual queries, actual data. Pick a specific instance, which
> I'll say is "foo".
>
> up{instance="foo"} # show graph over 1 week
> up{instance="foo"}[2d] # show console (range vectors can't be graphed)
> max_over_time{instance="foo"}[2d]) # show graph over 1 week
>
> I assert that if up{instance="foo"} is a mixture of 0s and 1s for a given
> 48 hour period, then the value of max_over_time{instance="foo"}[2d]} at the
> end of that 48 hour period will be 1.
>
> On Tuesday, 16 August 2022 at 15:08:15 UTC+1 [email protected] wrote:
>
>> hello,
>>
>> max_over_time(up[2d]) == 0 is giving me the info like ...for the last two
>> days if the server goes down for 1 minute also it was displaying in the
>> graph which I don't want. I want the information that for the last "X" days
>> it should be completely in an unreachable state.
>>
>> Thanks & regards,
>> Bharath Kumar.
>>
>> On Tuesday, 16 August 2022 at 12:47:03 UTC+5:30 Brian Candler wrote:
>>
>>> If the metric is 0, 1, 0, 1, 1, 0 ... then max_over_time will be 1, if
>>> the time period in question covers those values.
>>> If the metric is 0, 0, 0, 0, 0, 0 ... then max_over_time will be 0.
>>>
>>> If you enter an expression like
>>>
>>> max_over_time(up{instance=~"some_instance_name"}[2d])
>>>
>>> and *draw a graph of it*, then you need to understand what that graph
>>> represents. On the X axis is time; this is the time the expression was
>>> evaluated at. The expression itself looks at the 2 days of data *up to and
>>> including that time*: that is, the range vector up[2d] reads all data in
>>> the database between T and T-2d.
>>>
>>> For example, if there's a point on the graph where the X axis is 15 Aug
>>> 12:00, and the Y axis is 1, it means that the max_over_time between 13 Aug
>>> 12:00 and 15 Aug 12:00 was 1. This in turn implies that there was at least
>>> one 1 value in that 2d period. It will only show 0 if *all* the values in
>>> that period were 0.
>>>
>>> If that doesn't do what you want, then you'll have to describe exactly
>>> what you see more clearly, with actual concrete queries and responses, and
>>> explain why it is different to what you expect. Otherwise, only you can
>>> see the data in front of you, so it's up to you to understand why your
>>> query isn't doing what you expect.
>>>
>>> > But I want only unreachable state servers over a period of time?
>>>
>>> That will be those where max_over_time(...) is zero, and you can filter
>>> down to just those servers with an expression like this:
>>>
>>> max_over_time(up[2d]) == 0
>>>
>>> If you graph this expression, then all the data points will be zeros,
>>> but the points will appear and disappear over time. They will be present
>>> at time T only if all the values in the period T-2d to T were 0. If that's
>>> not the case, then the point will not be displayed.
>>>
>>
--
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/76ce5af4-c50d-4496-958f-e849cdb10afbn%40googlegroups.com.
