Something to worry about

🛎️ Get Help
1

@bobbyD

My C-7 hub is a bit over 1 year old.
For the first time and without any apparent reason the entire Zigbee network
went offline. I was able to restore everything by shutting down hub followed
by physically recycling power.
So far so good, hub is back on line, but my questions are:

  • Is this something to worry about or occasionally this may happened?
  • Should I check something else?
  • Should I simply forget about this case, nothing to worry about?

I have Hub Info driver installed.
Hub temp is 40C, Free memory is 372420

PS.
Hub is under Hub Protect plus I have an unboxed spare C-7.
I bought this C-7 in a rush when I migrated from C-5 to C-7.
Right after migration my Zwave network went crazy. But it was
due to few Zwave ghosts devices plus few incompatible with C-7
Zwave devices.

Blockquote

1 Like
2

@vitaliy_kh Zigbee radio going offline is a sign of an overwhelmed mesh. You may have a device going bad that is effecting everything else. Keep an eye on your logs to see what is going on.

3 Likes
3

Well, it also happens under low memory conditions. Zigbee radio is first to go, for some reason. Install @thebearmay’s excellent Hub Info driver to monitor free memory state. Installable using Hubitat Package Manager (“HPM”) from Dominick Meglio (@dman2306), another excellent tool.

9 Likes
4

I checked Zigbee logs (well, this is after the reboot plus power cycling).
The zigbee activity is next to nothing.
Or should I check any other logs?

5

Glad that it came back online. I personally never experienced this on any of my hubs, but have seen hubs under stress dropping the Zigbee to preserve resources. Both @rlithgow1 and @672southmain highlighted the reasons why a hub might be under stress. I've also seen one offs, with Zigbee dropping once and never again.

I would keep an eye on the hub and occasionally screen the Logs for warnings and errors or hyperactivity from multiple devices simultaneously.

4 Likes
6

This driver already installed.
Free m emry is 372420, temp is 40C.
But this was right after the reboot plus power cycling.

7

Of course I will follow this advice.
Should I create any monitoring rules with notification or maybe there is
already hub helth app?

8

I hope, this is my case as well.

9

No, you'll know if the radio goes offline again, as nothing will work. Screening past logs after the fact is just as good.

10

Like @bobbyD said no real need to, but since you already have the Hub Information Driver installed, you could consider setting up rules to notify on consistent high cpuLoad (maybe > 3 for 10min), high temp (>60°C for over 10min), memory below 120000, or the zigbee status becoming disabled.

7 Likes
11

Thanks, this is a good idea.
I will add these monitoring rules.
And thank you on threshold nubers.
Creating rules is not a big deal but what numbers to use is
very good to know.

6 Likes
12

Those should give you a little buffer, but are a good starting point. Recent memory management improvements should allow you to get down to 80000, and I’ve seen some hubs that consistently run a cpuLoad around 3+, but based on my C7s the numbers above are where I’d start.

2 Likes
13

Have you recently installed any new devices or apps or created new rules or edited rules?

14
  • Did not install any physical Zigbee and/or ZWave Devices.
  • Added support for Home Assistance (installed Home Assistance Device Bridge driver).
  • Added number of new rules related to the Home Assistance integration.
  • Updating the existing rules is regular routines.

I guess, all this are normal activities.

Zigbee Radio cutoff happened in a middle of normal hub activity.
All latest modifications were done couple days ago.

15

As "normal" as Zigbee going offline :slight_smile: Updating or creating new rules, unless you accidently set up a continuous loop, are unlikely to cause problems. Now, on the other hand, adding new integrations or new devices could have unintended consequences, including, but not limited to killing your radio :wink:

16

Did you install maker API? How many calls are hitting it?

17

No, I don't have maker API installed and used.
Integration with HA is done via Home Assistant Device Bridge designed
by @tomw and few other power users. I have no idea what is used on top
of TCP/IP for the HE-HA communication but status reporting from the
HA Device to the HE is near instant. Nothing needs to be installed on the
HA side. The only requirement - HA Device must be one of the supported
by HE driver.

18

So, this is something actually expected?

I 100%+ agree with this statement.
I am trying to use only trusted apps and drivers.
An if I am not positive about outcome I am testing things on my spare C-5 before
installing it on work horse C-7. The major difference however, C-5 does not
have any physical devices at all.

19

No, forgot the quotes around "normal" to emphasize abnormality.

Adding support for a new integration, is a major event, and not a routine (normal) activity.

When I plan to add new devices or integrations, I introduce them gradually and observe their impact on the existing system very carefully. I call it the "fish tank method".

If you have a well established fish tank, introducing new plants, fish, etc can have devastating effects on the existing environment, that's why you have to isolate and gradually make changes.

2 Likes
20

Home Assistant Device Bridge uses the HE websocket interface for an always-on socket to HA.