Let's talk about a set up that "self fixes"

That is an amazing amount of hubs!! :bowing_man:

1 Like

Hubitat runs everything.

I was thinking about how sometimes things get hung up and need a refresh/initialize. I suppose things like watchdog would be useful for that.

I am lucky in that, unless I let a battery go flat, nothing acts up.
I have been lucky in device choices.

The other thing you can do, again not necessarily self-healing but having some redundant sensors can help with things too.. for example I mounted a contact sensor on my bulldog water valve and it confirms that the lever actually rotated when I set it on or off..

I have been using these apps to notify me if my sensors stop reporting in (typically, because their batteries are dead):

Very useful!


As much as I have tried, I cannot get my wife to self-correct. :man_shrugging:


Why? I have 258 devices running on a single hub. What situation requires 8 hubs???


Is that 258 "real" devices, or a lot of "virtual" ones?

I guess it depends on what you mean by "virtual" devices vs. "real"? If you mean virtual as in "a group", I have none. If virtual, for example means TCP/IP devices that use "virtual" driver then yeah many are virtual. I have Denon AVR receivers, Denon HEOS speakers, Ecobee thermostats, an AlarmDecoder based alarm system, a WeatherFlow Tempest, an Ecowitt hub, 2 BOND Hubs, a Rheem Water Heater, etc. Honestly those virtual devices are more intensive on hub since many require frequent polling. So I feel like I tax my hub pretty hard, yet still can't see the reason to have 8 hubs. I haven't yet seen a reason to have 2!

For physical devices:

  • 12 Zigbee devices
  • 102 Z-Wave Plus devices

All RM, one does outdoor lighting, another does indoor lighting etc.
Many sensors, all but one Hub C4's.
Load "spreading"

Hmm ok. Guess my point is, even with 250+ devices, I have yet to have a reason to spread the load. My entire home is automated. Every single light, every switch, 50% of the outlets, cameras, motion sensors, door sensors, window sensors, environmental controls, etc. One hub seems to manage it all just fine. To each his own I guess :man_shrugging:


How many?

Spread out over several acres, multiple out buildings.
Farthest out devices are 200' from central location.
Very strong mesh(s). Many repeating devices.
Many, many intertwining rules.

1 Like

I have 4 Production hubs and a 5th as development. (two C5 and three C7) and the single reason for "Why?" is "load sharing by area". I want to stay below 65 Z-Devices per Hub. I want all my Internet Facing / Polling devices on a hub.

I'm not finding any difficulty in imagining 8 hubs over the physical area described. :slight_smile:

1 Like

My philosophy has been the idea of multiple meshes = interference. So I want to avoid that whenever possible.

I think this is the best way to describe what I am doing.

I think there's that possibility. However, Z-Radio packets are tiny and there are giant gaps between the bursts of legitimate data. In other words, the possibility of interference is approximately equal to the possibility of no interference. The RF layer of the protocol takes care of those moments of overlap, which effectively increases the possibility of no interference. (Listen before talk + Random backoff.) All of which must be balanced against the known bottleneck of the Z-Radio queue. Gig speed processors knock, knock, knocking on a Z-Radio door of 250k max and slower quite often. Parallelism wins in that balance, I find.

Another way of viewing it is gig speed processors knocking on multiple 250k radio doors. "Two doors" probably means a significant gain over a single door, but not 2x as your interference model suggests. "Four doors" means yet another significant gain over a single door, but not 4x. But what if it's "only" a 3.5x gain? Or (shudder) a 2x gain? Clearly it's still a gain. I get more Z-Radio packets per second, functional ones, from my set of 3 hubs than you can from one hub.

Said yet a third way, I think the interference of multiple Z-Radios is managed within the protocol such that there's no detrimental interference. 60 ZWave devices and one hub are all competing for a single radio channel. Interference is mandatory. Doubling the number of controllers to 2 for the same 60 ZWave devices does not exponentially increase interference. It's just 1/60th more.

1 Like

Maybe. Anecdotal, but I tried multiple zwave hubs. Made things MUCH worse for me. Ymmv.

1 Like

I have had very good success with 2 hubs by location (never tried 8 though) - one for the main floor the other for the 2nd floor of our house - since moving to the C-7 I'm still using multiple hubs but these are now by function not location (Z-Wave/Zigbee/Network). Also have used multiple hubs by location for clients as well - it's a great way to extend the coverage over larger and/or interference prone areas while reducing individual hub overhead.

1 Like