Hi Brian,
         I'm attaching responses for the previously mentioned queries. The 
only differences between query1 and query2 are start and end timestamps.
Lets say for a particular timestamp *1731304830 *(present in both the json 
files)

*Query1 (5 seconds window) *returns a value *764.0958754210372*
*Query2 (60 seconds window) *returns a value *42.8*

As per my use case, regardless of the start and end timestamp, the value 
returned should be the same, as my step frequency is 1 second in both cases
Setting *max_source_resolution=0s *also does not seem to help.
Thank you,
Kishore. 


On Wednesday, November 13, 2024 at 1:02:39 AM UTC+5:30 Brian Candler wrote:

> Can you show what exactly is different in the output between the two 
> cases, that you don't expect to be different?
>
> On Tuesday 12 November 2024 at 17:13:26 UTC Kishore Kumar wrote:
>
>> Hello Prometheus Community,
>>           I hope the person reading this has an awesome day, and thanks 
>> for helping out.
>>
>> I have a thanos query API endpoint, which if I query for a smaller start 
>> and end timestamp, it is giving a different value compared to the same 
>> query which has a bigger difference between start and end timestamps. I 
>> request to know what might be causing this unexpected behaviour? 
>>
>> 1) *Query with 6 second difference between start and end timestamp: *
>>
>> *<endpoint>/thanosquery/api/v1/query_range?query=sum(increase(envoy_cluster_upstream_rq*
>> *{*
>> *kubernetes_namespace="sample_namespacce",*
>> *kubernetes_pod_name="sample_pod_name",*
>> *envoy_cluster_name="sample_cluster_name",*
>> *}[3m])/3) *
>> *by (kubernetes_namespace,envoy_cluster_name)*
>> *&start=1731304828&end=1731304833&step=1s*
>>
>> 2) *Query with 2 minutes difference between start and end timestamp: *
>>
>> *<endpoint>/thanosquery/api/v1/query_range?query=sum(increase(envoy_cluster_upstream_rq*
>> *{*
>> *kubernetes_namespace="sample_namespacce",*
>> *kubernetes_pod_name="sample_pod_name",*
>> *envoy_cluster_name="sample_cluster_name",*
>> *}[3m])/3) *
>> *by (kubernetes_namespace,envoy_cluster_name)*
>> *&start=1731304800&end=1731304860&step=1s*
>>
>> Thank you for your time, 
>> Kishore. 
>>
>

-- 
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 visit 
https://groups.google.com/d/msgid/prometheus-users/5b798934-5767-46f8-9918-428c3b50f25en%40googlegroups.com.

Reply via email to