I have a bit of a challenge, and I’m not sure if it’s a bug, or what I’m looking at here. I’m hoping someone can explain this, or point me in a useful direction...
I’ve been chasing an issue where some device status’ are randomly not sync’d with the dashboard after rule machine activities. However, now I think it might be deeper, but have no idea where to start looking.
This happens with random devices, but all in question are GE zwave plus (not sure it matters, but most of my stuff is so I have limited other reference points). I have a rule that shuts a bunch of things off at 12:01AM, but some of those things show on at odd dimmer levels in the dashboard post shut-off. Note the “front banister lights” below...
Now note the events for the module ... it shuts off, and about 3 tenths of a second later, it immediately turns back on and goes to 92% (sometimes this indicates 32, or 61%), but the physical module never responds to the ‘on’ and stays off - so now the Hubitat thinks the impacted module is on at X% despite the fact that it is actually off. As expected, the dashboard reflects what Hubitat thinks.
Where are these echos coming from? I’ve searched all the rules, I have no schedule conflicts, and this issue seems like it might be unique to Hubitat. These exact modules all recently moved from Smartthings, and ST was never once confused about device state on any of them. This seems to randomly impact different modules on Hubitat every night, but the event history is always the same pattern for affected modules... it’s like there is an occasional echo or reflection in the zwave mesh that the hub is seeing/logging despite the module being smart enough to ignore it.
As it does every night, it happened again last night ... one module last night, but not one that was impacted the night before. This time the ‘off’, and then an ‘on’ 5/100ths of a second later (see screen shot)
Something that has me perplexed though, is that the rule is set to trigger at 12:01 ... but the logs seem to show all the actions starting at 12:00, and nothing at 12:01.
This is a daily thing for me ... sometimes 3 or 4 devices get bounced back on and show the wrong state. I looked at that thread, the generic device drivers don’t have a transition time to tune (not even sure what it would do)
I’ve never seen this occur outside of RM ... and at that, it’s random as to which devices are affected.
I haven’t tried using Simple lighting or an other method to get the job done, only rule machine so far. I’m open to giving something else a try though.
And then the plot thickens with this one ... the ‘reflection’ occurred roughly 5 minutes later. This one actually turned back on, so the reported status was accurate, but it never should have turned on. WTF...
Although this is more of a workaround than a solution ... I ended up creating a rule that runs a custom action and a refresh for all the devices in the original rule which runs five minutes after the original rule. This seems to update Hubitats awareness to the actual state of the devices, and doesn’t stomp on any intended states (if someone turned a light back on right after the blanket rule had previously shut it off).
I’m still at a loss about where the reflection came from, but at least the dashboard stays in sync after RM actions now ... outside of that five minute window.
