This query should be fast also for sensors with a million beliefs.
It was once designed to be (using order by and LIMIT 1) but that can easily break, given our unstable index.
(maybe #2378 will improve this, or SeitaBV/timely-beliefs#237 already has on main - I know there has been recent work on improving queries)
Anyway, with FlexMeasures 0.33 on our live system, this is the current waiting times on an asset with a lot of data (our asset 1):
Let's re-do this test when we run 1.0, and discuss what can be better.
This query should be fast also for sensors with a million beliefs.
It was once designed to be (using
order byandLIMIT 1) but that can easily break, given our unstable index.(maybe #2378 will improve this, or SeitaBV/timely-beliefs#237 already has on main - I know there has been recent work on improving queries)
Anyway, with FlexMeasures 0.33 on our live system, this is the current waiting times on an asset with a lot of data (our asset 1):
Let's re-do this test when we run 1.0, and discuss what can be better.