Device Health Reliability plummeted [September 2018]


(Bob Kerr) #1

After a great run of months with no false positives on device health notifications, in the past few days the reliability plummeted. Zigbee and Z-Wave devices are reported incorrectly as offline while they still work as expected. Is anyone else experiencing the same issues in the last few days?


(Larry) #2

Same thing, no logging,.no device status coming back, and their website show all normal . I think they stopped monitoring anything…


(DavidK) #3

Yes, I see the same thing.

I wonder if this is an app issue. So for either pairing or unpairing, status never gets to the iOS app, but can see on the server logs that I had a successful pair or exclusion.

Also, in the past when a device was listed as not available, turning it on or off would then mark it back as available.

Now turning the device on and off does not update the device health status.


(DavidK) #4

Yes live logging tab of graph website never shows any info.


(DavidK) #5

This is the support ticket that I submitted yesterday Saturday around 8:30 pm
#622008


(Bob A) #6

I am seeing the same, incorrect status of devices offline and live logging not logging anything at all. I just spent the past hour thinking it was something wrong on my end.


(Bob Kerr) #7

I like how Alexa turns the light on but then says the device isn’t responding.


(Jimmy) #8

Yeah something isn’t right. One of my devices marked offline is my front door lock. Check interval is 3600 seconds in the ide (60 minutes). Yet the device says it had activity two minutes ago. Well within 60 minutes.


(Bob Kerr) #9

Same with a bunch of my GE switches. Interesting enough, Z-Wave repair doesn’t think they are failed.


(Hendre) #10

My (Yale) lock also intermittently shows as being offline the past 3 days.


(Bob Kerr) #11

Interesting enough, none of the “offline” devices showing in ide/classic are offline in the new app. @prjct92eh2 ?


(DavidK) #12

@HA_fanatic

I noticed this also. Does the new app even have device health status?

I am not that familiar with the new app. Just noticed that there are no red dots or any indication for the same devices that show offline in the classic app.


(Allan) #13

Same here. About 4 GE Z wave light switches that all came back with a “refresh” button press and a couple ZigBee leak sensors that didn’t and haven’t manually intervened yet.


(Larry) #14

Everything seems to be working again for me


(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy) #15

I don’t understand why anyone uses “Device Health”. I have yet to read any posts anywhere that say it is helpful…

My system is 100% stable (i.e., no more outages or response issues than usual). There was a brief outage last week, and a brief period when response to triggers (e.g., contact or motion turning on a light) was “many” seconds, compared to “few or less” seconds.


(Bob Kerr) #16

Is not about using or not using the device health. I could care less if it says that device is unavailable in the app (I rarely use the mobile app anyway) although it’s a nice feature to have. The problem is what gets marked as “offline” in ide, which other services rely on, like Alexa for example - who wants to hear that device is unavailable because it was incorrectly marked as offline…


(Bob Kerr) #17

Yes it does and it used to match what was marked offline in the Classic app/ide.


(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy) #18

So… forgive my ignorance… (I’m on a Hub V1 and never have any offline devices)…

Isn’t this “offline marking behavior” disabled when you disable “Device Health Check”?


(Bob Kerr) #19

Ha! You might be right on that. I guess I never had a need to disable it before because it was just working. Fortunately, I cannot confirm your theory because I woke up this morning and magically had no “offline” devices. The next time this happens I will sure be testing this :wink:


(Larry) #20

I was getting device not responding errors in alexa not due to device health, but due to the bug over the weekend where no devices were updating status correctly