That was weird. I had a Samsung Water Leak sensor (one of the later ones with sensors also on top) get wet. I had a Basic Rule that triggered properly ...
When Water Leak Master Bath water leak detected ... Turn on * Alexa says Water Leak Master Bath
However, this same leak sensor is listed in HSM Water monitor, and Water was armed. But none of the actions for a water alert in HSM executed, leaving the main water valve open and no notifications. There were no events in the log related to HSM at the time. HSM failed me, fortunately with no bad consequences since I was home, and the "leak" was just the cleaning lady mopping the floor.
But wait, there's more. After this, I wiped off the moisture on the leak sensor, but the sensor didn't report dry (at least Hubitat didn't reflect that it had). I made really sure that the sensor was completely dry inside and out. I replaced the battery. And the sensor was correctly reporting temperature and battery events .. just not dry. So then I replaced the device type of Generic Zigbee Moisture Sensor with Virtual Moisture Sensor and saved. With this driver, I was able to force the dry state. Then I changed the driver back and it shows dry now (and I know that the sensor is sending events).
It definitely reported wet, as I have a Basic Rule that it triggered at the time (the basic rule causes Alexa to announce that there is a leak). That's how I knew it happened. It appears that it never reported dry again. But that wouldn't have kept HSM from alerting.
And I'm totally sure this water leak sensor is included and water was armed at the time.
There is the mystery of the device never went dry again and that may be a clue. But the fact remains that the device did raise a wet event that Hubitat saw, and HSM didn't take any actions.
What logging in HSM do I need to enable? The device does show a wet event at the time. There was nothing in the logs related to HSM.
FWIW, I just added a water sensor to my system, and setup HSM to alert me when it is wet. Test worked, I was alerted, then sensor reported dry shortly after.
I can understand why you want to rule out certain things, but a lot of this was covered in my original post. Yes, HSM was armed for water. Here are the HSM events at the time the incident occurred.
There are no event in the location events during the hour before or after the sensor going wet. Here is the other log though that shows the wet event at 11:53:08.197:
I can try to repro this tomorrow. It probably won't though .. I always test leak sensors every time I do a battery change and HSM and the Basic Rule always have worked in the past.
Well, as I expected, I was not able to repro HSM missing the wet event. I guess we'll classify this issue as one of those where if it happens again, we'll know it has happened before, and that's about all that can be said.
Possibly completely coincidental but the leak sensor is wonky. It reports temperature, battery, and wet perfectly. It no longer reports dry. Refresh, Configure, power cycle, and re-pairing do not restore its ability to report wet. The only thing I can do is to fool Hubitat into assuming it is dry by changing its driver to a virtual one, then changing it back.
I'm sure I'll replace it, but ironically the sensor still does its most important job just fine .. detecting leaks so it isn't going into the trash can quite yet.
This is a serious problem, because the platform would discard a wet event that was a repeat -- as it would not represent a state change. Without the intervening dry event, the sensor is worthless.
Understood. But I have forced it dry and put it back in service, but only until I can get it replaced. The platform thinks it is dry now, so it should work .. once. But that’s better than nothing.
This problem has occurred again, but without the pesky nuance that the one sensor was not following up with a report of dry (that sensor was discarded back in November). Also, this time I could repro it at will.
I accidentally wet a water leak sensor this morning and HSM, though Armed Water, did not take the configured actions. I reproduced at will with this and a different water leak sensor as well (one was Samsung, on was Zooz). Here's what I've verified so far:
HSM shows Armed Water and not currently in an alert state
Actions are configured including closing a valve, notifications, and light alerts. None of these actions are taken.
Both water leak sensors are configured under Select Water Sensors and show in Settings.
HSM shows no events related to the water leaks. It is just showing recent arm/disarm of intrusion.
The devices are definitely reporting wet/dry. Those are shown in device events and logs.
I have a rule in RM to take an additional action (Alexa announcement) when any sensor reports wet. It is being triggered as expected.
HSM is shown in "In Use By" for the sensors.
I then went back in the Water configuration, made/unmade a change to dirty the page, and hit Done. After this, it started working again.
I know this speaks to the importance of periodically testing HSM, but it is still disconcerting to think it can become unresponsive without any indication. I already use Watchdog to make sure the leak detectors are reporting regularly, but I can't think of any passive way to detect if HSM isn't working.
