Original timestamp when downsampling

[pierredfc]

Hello,

I'm trying to retrieve min, max and avg metrics from a defined interval (24h in my case). In order to retrieve min, I post the following request :
{"start":1557266400000, "end": 1557352799000, "queries" : [{"aggregator":"min", "metric":"sys.idle", "downsample":"24h-min", "tags": {"host":"MYHOST"}}]}'

The response is :
[{"metric":"sys.idle","tags":{"host":"MYHOST"},"aggregateTags":[],"dps":{"1557273600":5.760000228881836}}]

But the timestamp 1557273600 is not the timestamp when occured the value 5.760000228881836. I always have the day when occured the value following by 2AM (02:00:00).

So is there a way to retrieve the "original" timestamp ?

I saw on the documentation that the time stamps is compute with the following rule :

For early versions of OpenTSDB, the actual time stamps for the new data points will be an average of the time stamps for each data point in the time span. As of 2.1 and later, the timestamp for each point is aligned to the start of a time bucket based on a modulo of the current time and the downsample interval.

But can we override this behavior ?

Thanks!