I have had an issue with a switch only at night for weeks now just to find out my whole zwave network was crashing Oddly it affected me the most at night when the bathroom light switch wouldn't come on. Little did I notice other small malfunctions that were because of it. Mostly because it rebooted and works again on its own. Had this been an alert (on the bell in the corner) I would have noticed immediately. I had to dig through logs to find it. Can we make this an actual error alert? It would have been beyond helpful and it seems to be a critical error that should be noted.
EDIT: I have it fully monitored now with 2 lines of webcore code. That isn't the point. The point is this seems like a critical error that should be immediately reported.
See like a valid idea. I thought that when I was experiencing Z-Wave crashes, there was an alert in the hub UI. I guess you're asking for the alert to persist after a reboot?
In the interim, you could set up your own notification for the event. This is a rule I set up a while ago when backups were sometimes causing Z-Wave crashes.
Have you considered that it is that light switch causing Zwave to crash, and not that the switch stopping is just a symptom of something else?
The Zwave mesh seems very susceptible to going down based on even one "bad" or misbehaving device. You can see that in many of the Zwave issue posts here.
You may want to monitor the device with a rule, to see when the device status goes to INACTIVE state. I can't see how to do it in Rule Machine, but in Webcore you can check $status for changing to INACTIVE
That switch is currently brand new. I swapped it out and it still has the same issue. That's how I figured out eventually it was a network wide Issue. I am currently tracking down the bad device. I don't think zwave devices report in that they are active and alive like zigbee devices do.
I read that to mean the hub had already been rebooted before you looked for the alert. It makes sense that when the issue is resolved, the alert would clear. If you looked at the UI while Z-Wave was still crashed and didn't see an alert, then something is up, because I've seen an alert displayed there when my Z-Wave has crashed.
It sounds like you want to know that Z-Wave has crashed, even if it has already been resolved, without having to look at the location event logs. Creating a rule to send an alert when a crash happens would accomplish that.
I do have the hub info driver installed, but I'm pretty sure the location event I used in the rule I posted above is directly from the hub, not the attribute from hub info.
Yup, you are correct. Even pulling the air gap still reports ACTIVE for zwave.
When I had a bad devices crashing my Zwave, it turned out to be the devices that that had stopped responding to commands at times on their own, so your switch looked like a culprit based on original post that didn't mention it was already replaced.
In my case, the crashes seem to have been happening because my bad devices were not repeating when they went down, even though other devices were still trying to use them as repeaters.
Since that switch is still acting up but is now assumed "good" by replacement, it might be losing a path through a neighbor device when it stops working, which could explain why a new switch in the same physical location is still having issues with responding, since it still has the same neighbors.
I would look at the device graph and see what devices that switch tends to repeat through, and then suspect one of those devices.
I removed an old door sensor and removed and pulled the gap on the relocated old bathroom light switch. Everything is humming along happily now, but only time will tell. I'll need to keep an eye on it. I will keep an eye on the zwave device info page.
