Even after a prolonged time with no actions I do observe longer timesteps duringwhicg no data is put into the database (more then a hour, even with changes in the state).
Recommending only “everyChange” as strategy seems tobe not so good
OK, so does that mean we need to keep the recommendations in the docs? Maybe every 5 minutes or every hour would be better?
I have to admit that I was wrong! I have no real clue why that isn’t working.
As for the suggested minimal timestep for persisting I’m not sure. The second archive of the default datasourdes does keep a value every minute, I beleave that this timeframe is the most looked at, hence I would argue for everyMinute (although I argued against that before ).
If I understood you correctly, you evaluated the everyChange strategy, and it didn’t work reliably. Maybe report the issue somewhere? I see the potential for a lot of users spending a lot of time trying to figure out why it isn’t working.
If users read the docs as they are currently written, they are told in at least two places that an everyMinute strategy is required and if you don’t set a strategy yourself, the default strategy includes everyMinute.
What @opus and I were exploring was whether or not that is still necessary which it appears is the case. So we will leave these warnings in the docs and leave the default strategy alone.
In short, we discovered the status quo is still in effect, we didn’t discover something new.