How does Hubitat Safety Monitor handle a lost connection to critical device while armed?


#1

I have looked through the forum and documentation and have not come across any information on how Hubitat handles a lost connection to a critical device while armed.

There should be a communication heartbeat that HE is monitoring to make sure the device is still communicating and able to send a change of state back if once occurs. If a series of updates are missed, HE should assume that the device is no longer able to report its status back which should then result in the same action as if the "door opened" or whatever the device is.

In the Industrial Automation Safety world, this is done with normally closed circuits for hardwired devices and specifically requested packet intervals (RPIs) for Ethernet/IP devices that ensure the device is ready and able to tell you what it needs to tell you. Loss of communication is equal to the sensing of the unsafe action.

This all came about when my Smartthings multipurpose sensor that I am using on my garage door stopped communicating, luckily it thought the door was open and I did not arm my system for the night. What if it was the other way around? Would Hubitat let me arm my system with a non-communicating sensor?


#2

I found a good whitepaper talking about Z-Wave S2 and how it handles this. I would still like more detail as to how this is implemented in Hubitat.

I hope Zigbee has a similar architecture too!!


#3

Well, maybe Zigbee isn't as good....?

This paper seems to suggest that none of the devices they tested had a heartbeat.


#4

SmartThings had a feature called Device Health which had good intentions but was very error prone. This feature has been requested for Hubitat but staff has mentioned they want to get it right before ever releasing it. You can read more here:

There are community apps like Device Watch that may help you in the mean time:


#5

I think it would take a while before we have wireless devices with a heartbeat that are stable. Well the device would be, but the mesh is the issue. I would never bet my life on wireless. My plan is to extend all my doors and windows with a z-wave contact sensor besides the zigbee ones I already have. But if I ever move to another house I'll go for wired everything. It makes everything so much easier. There are simple, non critical, things that can be wireless like switches and buttons. But all other things like motion, contact, etc have to have a wire.


#6

There is no "monitoring" in the sense that you mean it. Everything is event driven. HSM is waiting for events from devices. If a device being used drops off line, is stolen, has its battery die, gets zapped by lightning, etc. HSM has no way to know that happened. That device simply no longer sends any events.

Device Health is a chimera. It is only as good as the devices involved, and in general most of these consumer devices are not of high enough quality to support true device health. The downside of pretending to have device health is that it leads to a false sense of security that doesn't actually exist. We've had lengthy internal discussions about this issue for years. We are not comfortable pretending to offer Device Health at this time.


#7

The technology exists, but to bravenel's point below, it does not exist in $20 consumer sensors currently (and a battery-operated one at that).

I would go a step further and say the downside of pretending to have a "security system" is that it leads to a false sense of security that doesn't actually exist. It is very disappointing to see all the major hub and "security" system companies relying on a technology that in the industrial world would not even get you the lowest category of safety compliance.

Without device health, Safety Monitor is a chimera.

Don't get me wrong, I LOVE my Hubitat (and Hubitat is not marketed as a security system, that's just a little bonus), I am just taken aback that there seems to be an entire industry based on companies (Ring, Nest, SimpliSafe, etc.) selling systems based on this same technology, and what I am hearing is the technology is not fail-safe and therefore should not be used (and marketed) for this application. :confounded:


#8

Well, uhm, yeah. Because wireless can never be flawless. Ever. A wireless security system is for a burglar just as easily removed from the equation as a tripwire in your garden.


#9

People make money selling uneducated people lots and lots of things. Take any of those systems and call your Insurance company and ask for a "Security System" discount on your insurance and see what they say? There is an industry certification/qualification for security systems that must be met and those systems do not qualify. What you will notice is that "Real Alarm" systems will offer Z-Wave or ZigBee as additions to their alarm system. However the actual "Security" device part of the alarm ARE NOT Z-wave or Zigbee.


#10

It's a monitor of the system you have put in place. It's as weak as its weakest link. It is what it is.

Don't count on it to be an industrial strength security system. Is it better than nothing? Yes it is. Can it provide something useful for your home? Absolutely. It can warn you about a water leak, it can warn you about an intruder or a broken window. It can warn you about low temperatures.

Security systems are the chimera here. They don't keep you safe. If your house catches on fire and becomes fully involved, you lose your house no matter how good your smoke detectors are. If someone wants to burglarize your house, your security system is not going to stop them. Security systems are sold with a good dose of fear and unrealistic hope -- the guy breaking down the front door on the wife, and somehow the security system saves the day. None of that is real.


#11

You do not need flawless communication for it to be fail-safe. You just need the system to react the loss of a device in the same way as the device being tripped. "Removed from the equation" in a normally closed circuit is impossible as that will set off the alarm just like the device was there and operating normally.

Yes, it would be possible to get "false alerts" if you have poor signal strength to a device, but you should be able to configure how many heartbeat or polling update losses in a row results in the determination that the device has been lost. And yes, a "false alert" from a drop in communication would be annoying and scary but if you truly want to have a reasonable level of safety, that is what is required.


#12

I'm not a burglar but wouldn't it be very easy to then trip (by interference) the alarm a couple of days in a row without leaving any evidence of a burglary and remove all trustworthiness of the system and make people so irritated that they leave it off and then just do the heist?


#13

I like the "disclaimer" at the beginning :smile:

You are correct, but that would be an active decision from the user to ignore it and not a failure of the system to alert you of a problem. There are downsides to everything, but if you trust the design of your system and have the correct tools to determine what went wrong, why would you disable the system and not investigate why it happened?

I think it would be interesting to see if there is a way to have another level of devices like "safety-rated" devices that obviously cost more and go through batteries faster, but provide a reliable connection with minimal false alerts and have device monitoring as standard.

This is the way it is in the industrial automation space. You have cheap I/O that you can use for simple automation, but if you want I/O to be used in the "safety" system, you need to get specific hardware that supports the redundancy and fault detection needed for that level of MTTFd (Mean Time To Failure, dangerous).


#14

This already exists with real alarm systems.


#15

Right! Lets make a stronger "weakest link"!! A contact sensor that has dual-channel feedback (one channel over Z-wave and one over Zigbee) that way burglars would have to jam both frequencies at once to take the system down in a 1oo2 configuration.

I think there is a market for it since millions of people are buying the crappy versions (me included lol).


#16

Right, with $60 a month contracts and closed ecosystems.

That does not mean the open systems can not take the technology and apply it to continue to take market share :smile:


#17

I'm not talking about ADT or some telemarketing system that you get for "free" with their service contract.

The most capable and open system is ELK. Another that is commonly integrated is DSC.


#18

Never heard of either of them!! Will be looking into them. Thanks!!

It seems like there is large gap in the industry between super high priced "professional systems" and the cheap systems that cannot offer much reliability at all. Or maybe there are companies that fill that gap like the ones above, but no one has ever heard of them lol.

Thanks!!


#19

There's a gap... it's between consumer and professional.


#20

The reason you haven't heard of them is that you are not their target audience. The installation / security companies are.

Besides all that I am curious to how you live. Because I would normally use a simple rule for my security: always make sure it's easier to break in to you neighbours house. Now, I live a in a dense area.