A problem when applying a filter to heat maps

[M.alsioufi]

I have a field of type date and default format insertion_time
I wanted to create a heat map visualization in Kibana 6.5 where I can have and the day_of_month on x-Axis, the hour of the day on y-Axis and the color represents the count.
So I had to create scripted fields that extract the hour and day_of_month as numbers and store them as sub fields under insertion_time.
The scripts I used are:
doc['insertion_time'].date.dayOfMonth forinsertion_time.day_of_month field and
doc['recording_time'].date.hourOfDay forinsertion_time.hour field
In the visualization settings, for X-axis Aggregation field I chose Terms the field is insertion_time.day_of_month, 31 values and the order is Alphabetical and Descending.
However, whenever I apply a filter or try to split the chart on some fields the order breaks.

My questions are:

1. Is there any other way to handle the hour and day_of_month fields as date type? so Kibana can understand the way they should be ordered with?
2. How can I prevent the visualization from re-ordering the terms on the X and Y axis?
3. Are these issues solved in later versions of Kibana?
4. Is there any better visualization that I can use to show the data distribution I am interested in? Even if this requires installing some other Kibana plugin.

Thanks in advance..

[Nathan_Reese]

Use histogram instead of terms aggregation since the fields are numbers. This will give you the option to show all buckets.

Looks like the ordering is still a little wonky when a filter is applied. I filed https://github.com/elastic/kibana/issues/40790 to track.

[Nathan_Reese]

If you use histogram aggregation and set "show empty buckets" and configure "Extended bounds", then the heatmap will work as expected.