I am in Arizona, it is 108 deg right now and same yesterday. Not surprising for a power outage or brown out. But I have had two short term, 1 min or so, power outages two days in a row…
See my meter charts for the two days.
It appears there was no readings at all for a time period leading up to the outage, then it measures for a bit, then the outage occurs. Then comes back on…
What am I experiencing? With the no reading…
Hi Jim,
I was looking into the issue, and I don’t know why this would happen. One issue could be that Sense has to reprocess data from “Real-time,” which predicts what the power meter should look like based on “real-time” estimates. This is because Sense needs to display data faster than it can be processed. Then there is “Historical Data,” which occurs about 15 minutes later, where the data is more processed and accurate and could differ slightly from what was seen beforehand. It could have been lost if power had been cut while some data was being processed.
Generally, this is not the case, or if it is, there isn’t a little bit of data before the actual outage. However, users with power outages tend to be less specifically able to isolate them.
If you or anyone else sees more issues like this, let me know, and I can look into it. However, I would say generally, this is just a data loss event due to a power outage, where Sense also lost power.
I know this may not be super specific but I hope this helps.
Kinda get what you mean somewhat….but wouldn’t you think even the actual power loss wouldn’t be there?
Wondering, if maybe there’s some kind of a way to have a firmware cache, for events like this , so that there’s is some kind of recovery…(of course there would be time limits to this) and depends on hardware/architecture)…these were both less than a minute….
I just assumed it was how the system worked (and didn’t really question it, and, since outages are fairly uncommon, didn’t really have a huge impact on my overall numbers). My experience has been that a power outage on the Sense circuit (whether a whole-house outage or a flipped breaker on Sense) has me lose the 10 minutes (approximately) of data prior to the outage. I never noticed it before, but looking back on a very short outage from a few weeks ago, I also see the large gap, short set of data, short gap, then back to normal operation that Jim shows above. I’m not sure if the timing works out though, because my gap runs from 5:19AM to 5:28, then again for about 30 seconds in the 29th minute of the hour. My Nest Protects (hardwired, but with built-in battery backup) reported power loss at 5:25. The total outage was on the order of a few seconds during a particularly nasty storm.
If anyone at Sense wants access to my data to investigate further, go ahead. This was May 24 at around 5:30AM Central Time.
I have not done any testing to see what happens if, for instance, I am watching the power meter display, reset the breaker, and see if the data just disappears from the app (with or without refreshing the view). I just grabbed this screenshot today (June 10), so I think it’s safe to say the gap isn’t just a processing delay; I don’t think it’ll ever get filled in.
We save data if Sense goes offline, but power outages are trickier since the Sense is offline and not even powered on.
I can’t really speak for what should be done, but if data is lost and not recorded, it is generally not able to be recovered in the future. If there is evidence of a power outage around the time data is lost, it can usually be inferred that they are related, though it is possible they are not.
This is why it could be useful to see if issues like this were happening when there was no power outage as well, so we can investigate further. The data would not be recovered, but it could be an issue worth looking into.
Is there a way to hookup eg a 9 or 12 volt external (rechargable) battery to just power the sense unit over poweroutages and keep the data gathering alive as well as the un-send data to the cloud ?
Thanks for that info, thought it might just be me…but appears to be a behavior……Sense…could this have to do with how often you acquire data and/or how much at a time? Is there no internal, non-volatile cache to cover episodes of some duration. Mine were very short.
