Im getting some weird behaviour with the time since event that I just cant understand....
First of all when editing the condition with 'time since event' it doesnt display correctly - the 'time boxes - h:m:s' only appear if you fiddle with comparison box and attribute box. But this isnt major and is manageable.
As you can see in the screenshot it shows that time since event has NOT exceeded 1hour, and yet when I look at the device there have been no events since the start of December (as the device dropped off the network). So this should be a 'true' condition, but seemingly not.... I enabled logging on the rule, but everything looks sensible except for the fact that this condition is just incorrect....
From your screenshot, it looks like there is no humidity values being reported at all, hence the « ? » In « Bathroom Temp(?) ». I suspect there needs to be at least one event reported for this to work.
Thanks for the response! Hmm thats an interesting point, yes the device stopped reporting humidity and temperature measurements as it dropped off the network for some reason. There is a previous value stored (for humidity and temperature), but the event log doesnt have anything in it as it dropped off the network about a month ago. I added that condition to the rule to try and prevent mistaken behaviour when the device drops!
I guess this means the behaviour of the rule is a bit broken as it really means 'time since event' is greater than X and also less than event log size.... I.e. a 'null' event triggers rules irrespective of condition....
Doesnt fix the bug with displaying time fields, but at least I can try and find another condition that behaves properly.
This was fixed in latest release. So update. Another work around is to pick something other than **Any Value**, put the time in, then go back and select **Any Value**.
I feel like you've been pretty clear in the past about the risks/limitations associated with using TSE, especially when used over longer periods of time.
I'm wondering if there is a way in RM to identify when the current value of a TSE condition is "?". Here's my reason. For conditions like "TSE > X", RM works great until of course the TSE resolves to ?, at which point the condition is no longer true. If there were some way to trap for this, we could handle these edge cases.
I've also gone down the road of adding hub variables containing the actual time since key events have occurred, but that approach seems to have it's own set of tradeoffs.
