How does date histogram devide time to buckets?
Example: interval 110m
I know that If I have GMT timezone then I could start splitting time from Epoch (1970.01.01 GMT)
e.g. 1970.01.01 + 110m + 110m + 110m..
If I have timezone with positive offset (e.g. Europe/Moscow).. then I have to use Epoch 1970.01.01 +004
and calculate [1970.01.01 +004] + 110m + 110m + 110m...
but if I have timezone with negative offset like "America/Los_Angeles" then Epoch time is not working... at least I can't calculate correct buckets values in the way I have described in elastic search 2.4
Could anyone give me some explanation?