Due to the false physical events we were seeing, I pared my list of devices in Z-Wave Poller down to 6 absolutely critical devices. I added the rest of my non-plus devices to a series of RM rules that refresh one device per minute in a 30 minute loop.
Regarding this update:
Will this eliminate the false physical events we were seeing? I'm wondering if I should move my other non-plus devices back into Z-Wave Poller now. I am still seeing false physical events with the RM rules, but since the refresh frequency is significantly lower, so is the frequency of the false physical events.
@bravenel any input on this?
I think I'm not understanding how the RM Polling actions are supposed to work. I have Z-Wave Poller set to poll 6 devices:
This works as intended.
As a test, I also set up the following rule to poll a 7th device:
This device doesn't seem to be getting polled as is not getting updated. It doesn't show in the Poller logs (should it?) and the status never updates when I change the wall switch.
Here's the Poller logs:
Here's the RM logs for this rule:
I too would like to see what @bravenel and the team have to say about this. I'm still seeing the false physical events (mostly manifesting as on at 10%, but sometimes at strange % values) for polling.
Unfortunately, because of the Eaton Cooper non-plus switches I'm using, anytime I have an automation attached, I have to poll, as not doing so causes the automations to fail, normally because the automation state doesn't get updated when a light is turned off from that automation, and walking into a room with a Motion Lighting app attached to the light and sensor doesn't fire, as it thinks the light is still on.
Manually polling or refreshing the light in question resets things, but that obviously isn't a scalable solution.
So, while I hate polling as much as the next person, it would be good if we could get the Hubitat team to help figure out why polling is causing these false physical events. Thanks.
Also, and this may be related - I saw something interesting today on one of my sensor/light combinations (I've backed polling down to just 7 dimmer switches now, as the occurrence of issues seems to climb with more things being polled).
My sensor detected motion, and turned the light on, but only to about 10% (even though the automation says to turn it to 100%). I'm wondering if the polling for that switch hit at exactly the same time as the 'on to 100%' command hit?
Also, another thought I had is that Motion Lighting checks to see if the light is already on before issuing the on command (the 'on' parameter in the application state page), but I'm not seeing this issue in the one or two places I use RM instead of motion lighting. RM will always send the on command (which isn't a big deal even if the light is already on). I may switch a couple of Motion Lighting apps back over to RM to see if this alleviates the need for the polling to make the automation work.
I am experiencing false physical events on my GE switches. This only happens when the poller is on. Any tips?