The aggregate queries used to calculate cached aggregate values as well as serve live queries for non-cached aggregate values (e.g. sub-hour aggregates) can be slow to execute. The non-cached aggregate queries especially impact SolarQuery API performance. I've done some research and found the aggs_for_vecs extension can help with performance quite a bit, as well as using the Timescale time_bucket() function instead of the existing SQL function solardatm.minute_time_slot().

This ticket is to start with the non-cached aggregate queries, and re-write them using these better-performing functions.