Issues with statistics for sensor which is sum of sensors reset daily but not exactly at the same time

Hello,
I have quite specific issue with daily solar production. I have hoymiles microinverters reporting production daily for each panel. I get this via influxdb because that is how I made it done before (ioBroker gathering data from RS485toETH converter and pushing to influxdb). So far so good… but the problem is that every day I get strange statistics reading. After digging into details I found out something I actually new before - hoymiles resets values for each panel every night around 1am but not exactly at same moment. Each sensor will reset correctly but not at the same time. What I use for statistics is:

solar_total:
      friendly_name: 'Solar Production Daily'
      device_class: energy
      unit_of_measurement: Wh
      value_template: "{{ (states('sensor.Panel_SW_1')|float + states('sensor.Panel_SW_2')|float + states('sensor.Panel_SW_3')|float + states('sensor.Panel_SW_4')|float + states('sensor.Panel_W_1')|float + states('sensor.Panel_W_2')|float + states('sensor.Panel_W_3')|float + states('sensor.Panel_W_4')|float + states('sensor.Panel_S_1')|float + states('sensor.Panel_S_2')|float + states('sensor.Panel_S_3')|float + states('sensor.Panel_S_4')|float )|round(1)}}"

plus from customization:

sensor.solar_total:
  state_class: total_increasing
  device_class: energy

device_class: energy

So as result I get solar_total which does go to zero within 1-2min but not immediatelly drops from previous day production to zero. There are few values in between because each panel will drop to zero at different moment. Then I get some strange values for statistics. I am still suprised because generally value goes down and the configuration is “total_increasing” but I guess this is caused by the 10% margin mentioned in documentation (or something like this).

Anyway, I am thinking how to configure it. While writing this topic I just though of adding multiple panels separately to solar production. I guess this might solve the problem, if so - I will update this topic. But still would be good to find a solution for such non-typical behavior of total value (drop to zero but not at once, with few steps in between.

EDIT: Did not have to wait till next day after adding separate panels to energy dashboard (I guess statistics is available already). Yes - this solves the issue as each statistics is correct (each panel resets to zero properly). Anyway - maybe there is also a way to make it done for total sensor as described above?