Nik Cubrilovic

this will improve further with #88

Hi Christian - that is exactly what we're doing in V4 - when we first started using continuous aggregates with timescaledb ~2 years ago we found that they weren't flexible...

Oh - and we upgraded to SqlAlchemy 2, all async, param'd `text` queries etc. etc.

I'm re-opening this and tagging for v4 because there are improvements we can make for the v4 API After some reading earlier found timescale now supports real-time continuous aggregates: https://docs.timescale.com/use-timescale/latest/continuous-aggregates/real-time-aggregates/...

> You could just expand your existing dumps to cover the full history. It's there - we just haven't promoted it yet because we want to validate all the data....

Here's one answer: https://magic.io/blog/asyncpg-1m-rows-from-postgres-to-python/ 😮

> Sounds like you have the right ideas.

yep - got the alert but thanks for filing :) it corrected itself - we also have an ongoing issue with rooftop data from our source on WEM which I'm...

i'll leave this open to track the rooftop fill as well - it's an overall WEM data quality issue we want to eventually have the status of "data quality" on...

I won't force-regen since prod is a bit busy atm, but they're updated on prod + dev. The facilities list gets updated every 15 minutes anway.