Need to find the most actively reporting entity on my ZEN17 device

[clear chasm]

I may be using the wrong terminology in the title. Basically I would like to determine which attribute or entity of my ZEN17 is reporting to the controller the most often. Hopefully this will be something reporting every 20 seconds or less so I can use it for "watchdog" purposes in an Automation. What I mean by that is that is that I need my Automation to detect when a device goes offline (unplugged, dies, etc.)

Anyway, I've checked out the Log viewer and set to "silly" and was hoping to see a flood of activity, but in 2 minutes all I'm seeing is:

Subscribed to Z-Wave log messages…
Log level changed to: Silly
2026-04-21 19:29:46.150 DRIVER   one or more queues busy
2026-04-21 19:29:46.156 DRIVER » [REQ] [GetBackgroundRSSI]
2026-04-21 19:29:46.161 SERIAL » 0x0103003bc7                                                         (5 bytes)
2026-04-21 19:29:46.169 SERIAL « [ACK]                                                                   (0x06)
2026-04-21 19:29:46.177 SERIAL « 0x0107013b96939398cc                                                 (9 bytes)
2026-04-21 19:29:46.180 SERIAL » [ACK]                                                                   (0x06)
2026-04-21 19:29:46.183 DRIVER « [RES] [GetBackgroundRSSI]
                                   channel 0: -106 dBm
                                   channel 1: -109 dBm
                                   channel 2: -109 dBm
                                   channel 3: -104 dBm

Not sure if this means my Z-wave network is just really quiet or if the logs won't display the device updates.

I tried the Z-wave JS Logs viewer and can onyl trigger log entries when power cycling my ZEN17:

I'm getting the feeling that the ZEN17 doesn't call home as much as I had hoped it would. This of course is a good thing, in the grand scheme of things, but bad for my particular plan.

I asked in the Automations channel for ideas on implementing a watchdog feature and the detect-missing-device-updates-for-n-seconds (i.e., heartbeat monitor) idea seemed like a good one, but now I may need another approach.
https://discord.com/channels/330944238910963714/1496311399977062472

Anyone here have any ideas on how I can detect a Z-wave device going offline within a minute or less?

[balmy olive]

You can't without actively pinging it

Unless it's normal to be active. Then you can just use the last_seen sensor.

And pinging is only for mains-powered devices. Battery devices won't respond.

[azure marsh]

In Z-Wave, you don't want devices to phone home every couple of seconds. If just a handful of devices do this, your network can easily get flooded.

The same goes for pinging, polling, etc. Any automatic, repeated communication should wait for ~10 seconds of silence first.

[clear chasm]

10 seconds, even 20 seconds is acceptable for my use case. From a quick search it seems like zwave_js.ping is what I could use to initiate a ping.
When you and @balmy olive refer to using ping for my use case, are you thinking that I would ping and somehow await the response? Or is it more like ping will trigger a device status update and that is the data my automation should detect?
I think it’s the latter, but want to confirm.

[azure marsh]

Pinging is going to update the device state (alive/dead). I still don't recommend doing that too often. One ping of a single device every 10 seconds is about the maximum you should aim for. Less when there is other traffic

[clear chasm]

I agree, I don’t like to hammer a network or resources. The more I think about my use case the more I’m realizing I can lengthen the duration between pings. It’s not “The Running Man”; my dog won’t be at the line waiting for the fence signal to drop and then bolt for freedom!