Good evening, I have indeed recalculated the index from the beginning. I don’t really understand how the elements are prioritized.
@pierre-gilles and @Terdious
I’m checking in to see whether this development to recalculate consumption and costs in 30-minute intervals over a specific time period is mature enough to be integrated soon. If that’s the case, I’ll wait a bit longer to recalculate my data from the start of January only, to try to fix what went wrong for me on January 5th. That would allow me to avoid redoing the full calculation that took me about 30 hours in December 
But if there’s still work to be done before this improvement is available, I’ll rerun the full calculation… It’s not a big deal either 
Hi @StephaneB,
Sorry, it’s my fault, I have some corrections that @pierre-gilles asked me to make on the PR for Tasmota… And I won’t have time before next week…
1 Like
Whoa, it’s nobody’s ‹ fault › anyway! The project is moving at the pace of those who can contribute, no problem…
I just wanted to see if I should wait a bit because it was almost there. And so, since that’s not the case, I’ll take the other option, quite simply 
2 Likes
I restarted the full calculation, and it correctly fixed my abnormal consumption peak calculated on January 5.
However, I wanted to point out one thing, @pierre-gilles: the consumption calculation took about 12 hours, and the cost calculation about 15 hours. That’s similar to what I had the first time, no problem, I know my database is a bit large But the annoying thing is that Gladys was completely overloaded during the cost calculation: several very long minutes before displaying the dashboard, and execution of scenes was heavily slowed (about 15 minutes delay, see the example below)…
Does it make sense to you (and is it technically feasible) to lower the priority of these calculation tasks, compared to the operation of the dashboard and scenes?
My Gladys runs on a mini-PC, not very recent so probably not very powerful, but it’s not just a simple Raspberry
This scene with a trigger at 10:05 PM didn’t actually take place until 10:20 PM
1 Like
We could indeed reduce the rate, which would, however, lengthen the generation time — it’s a trade-off to find 
If we want to do that, we’d need to « estimate » the machine’s power to find a rate that suits it ^^
I think that in any case, @Terd
4 Likes
Hello,
I’ve had the same contract since this week.
Maybe a lead: