Strange behavior with bucket/pipeline aggregations

Hi
I have logs index with events logged every sec. The date filed is request_dtm.
I am trying to build the data visualization to plot in below format to have 24 rows per a day.
"Transaction Hour - Total Requests - Min TPS - Max TPS"

The Min TPS = The minimum TPS processed at any second in that particular hour.
The Max TPS = The maximum TPS processed at any second in that particular hour.
Total Requests = The total number of requests processed in that particular hour.

I have used the bucket aggregations with metrics
Metric 1 - Aggregation:Sum Bucket - Bucket:Date Histogram on request_dtm with interval as Second - Metric: Count
Metric 2 - Aggregation:Min Bucket - Bucket:Date Histogram on request_dtm with interval as Second - Metric: Count
Metric 3 - Aggregation:Max Bucket - Bucket:Date Histogram on request_dtm with interval as Second - Metric: Count

Split Rows: Aggregation: Date Histogram on request_dtm with internal as Hour

Problem Statement:
The output is bit inconsistent as the min, max TPS aggregation values are resulting inconstantly based on the time duration selected in Kibana.
For example:
1- See the first image with the time duration as for shorter hours. The min, max TPS values seem OK.

2- See the second image with the time duration increased. The min, max TPS values are changed and not same as resulted in Image 1.

I am bit confused on this inconsistent behavior. Can anyone please point at where I am probably making a mistake to meet my mentioned requirement.

Thanks

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.