Lag in device activity in the app

Has anyone experienced a lag in app reporting activity? When I first installed sense, it seemed that the app would give me a true live account of all activity in circuits. And with some appliances, it’s still does that. But I did notice with some devices, it’s completely out of sync, It would show my dryer turn on after it he completed its cycle. It’s driving me crazy because I’ve been trying to identify mystery devices and the best way to do that is a live view of something turning on and off. But i’m staring at my dryer in the off position and seeing a live draw in the sense app.

Which display(s) are you seeing the lag in ? The main Power Meter, the Bubbles, the Timeline and/or the device Power Meter for the dryer ? I could see the last three lagging because they all depend on detections, and some kinds of detections may no longer be immediate.

Which display(s) are you seeing the lag in ? The main Power Meter, the Bubbles, the Timeline and/or the device Power Meter for the dryer ?

The bubbles seem to be the biggest issue. When I check timeline, it’s actually correct. The bubble continues to be present while I know the appliance is turned off. The timeline also confirms it turned off, although at the top of the device details, it incorrectly shows the “on for xx min” counter

1 Like

I’ve started noticing many of the bubbles will remain long after a device turns off. For example right not my oven is off. I can confirm it is off, it shows “Oven” turned off under “Devices” “Device timeline”. However “Oven” still show’s it is on (green with wattage) under “Devices” and “Oven” bubble is still showing.

10:40pm -
image image

As soon as I posted this “Oven” bubble went away and is no longer green under “devices” , “Device Timeline” doesn’t show it turned off again. No idea if this qualifies as a “lag” or what the heck is going on but something isn’t right.



Sometimes Sense misses an off signature and bubbles stays on well after the device turns off. Sense will eventually timeout a stuck-on device after a period determined by the type of device.

1 Like

I’m familiar with devices missing the off signature and that is exactly what I thought had happened here. But this has an odd twist.

The Oven was only on from about 8:18-8:40. You can clearly see the cycle spikes in the devices “power meter” until 8:40pm. The timeline matches up to that point, Consistent on/off/on/off, green/grey/green/gray and so on.

At 8:40 “Device Timeline” shows “Oven (Kitchen) turned on” and didn’t turn off until 10:23pm.
But you can see Oven as still being on/green and using 466w at 10:40. The Bubble was also on the “Now” screen showing 466w. It stayed this way until after 11:00pm.
But if you look at the power meter is shows 0w from about 10:10pm on.

Again up until 8:40 everything appeared great. What happened after that I have no idea.

The device timeline, device power meter did not and does not match up to the bubble or the (non specific) devices screen.

Assuming the oven may have cycled on right before we turned it off at 8:40 (possibly a min or two later) and it missed the off signature. It appears Sense eventually did pick up on it but when and why the inconsistency how it is displayed?

10:10 when the Power Meter returns to 0W?
10:23 when the Device Timeline shows it last turned off?
11:10 when the bubble went away and the timeline turned the device to grey?

@paul.golota Is this similar to what you are experiencing with the bubble staying on long after the timeline shows it turned off?
BTW: there is a known bug regarding the graph section of the device screen Usage, Time on, Total times on hasn’t displayed properly since at least last year.

All good questions. I do know that there is plenty of “art and magic” in data science put into handling and managing data that is questionable or missing. And the arithmetic and aggregation surrounding that data can give different answers depending on what you are trying to do with the results. I wouldn’t expect Sense to have two different views of on/off for a device during the same period of time, but then again, we don’t really have a view of the underlying data representations - there might be states like “I think it’s on but not 100% sure”, “pretty sure it is off”, “I’m clueless”, and “I determined this was on, but well after it had actually turned on (for back filling slow ramp detections”). Calculating the number of times on and off, plus actual elapsed time on, isn’t black and white with those kinds of “unknown” states.

Here’s what I mean - imagine a calculation algorithm, that has an “If device is On, Then…Else…”. With more states that reflect various degrees of uncertainty, the same calculation might have either many branches of If, then, else, or a case statements, where the algorithm is tuned toward false pessimism removal (don’t make the calculation result in unknown / NA, unless forced to).

I wouldn’t either.

When a device is stuck on (misses off signature) normally everything matches up, a combined wrong of sorts.
For example I have a small under desk heater, that consistently misses the off signature, many times the on signature and correct wattage as well. But every time it misses the off signature the power meter still shows usage until the bubble goes away and the “Device timeline” records it turning off.

For lack of better words what happened with my oven appears to be an oddity in the normalcy of wrongs.

Your philosophy there could be additional dimensions of existence is quite intriguing. A seemingly logical possibility.
A ripple in time-space (timing, lag)…

@kevin1 our assumptions were wrong but we were on the right track.
There are two models, one saw an additional cycle, the other did not.
Check your PM’s.