I’ve had a very stable Zwave environment for several years. I use the Z-Wave JS (native) integration and currently at v. 15.22.1. That’s what I read from the logs during startup but I also see 1.1.0 on the “app” page so I’m a little confused.
My install is an OVA on VirtualBox.
Installation methodHome Assistant OS
Core2026.4.2
Supervisor2026.03.3
Operating System17.2
Frontend20260325.7
The Problem
About once every day or so, all of my Zwave devices go to “Unavailable”. I find the integration in “initializing…” status. I’ve found two ways to fix it: 1) Reboot the OVA (or maybe just restart HA) or 2) “Reconfigure” from the Integrations page and then restart the ZWave “App” / Integration.
I’ve turned the logging up on ZWave and see absolutely nothing out of the ordinary in the logs (except less traffic). I also see nothing in Core or other areas.
Any pointers on where to look? This started happening in the past 2-3 releases, it seems. (About the time we got the “UI” experience added)
device info >> 3dots next to configure >> statistics
Dropped RX: Messages that reached the controller but were discarded (often due to being malformed, duplicates, or checksum errors). High counts here can indicate interference or a dying node. Maybe check batteries for affected device. I change batteries at 40%. Lower and ive noticed issues.
Dropped TX: Messages the controller failed to send out. This usually happens when the network is too congested (high traffic) or the controller is busy. no good advice here but maybe check for chatty device. if you find it try rebuild route for the individual device.
Start by checking controller and based on what you find check individual devices.
I appreciate you taking the time to respond. This isn’t a network/mesh/device health issue.
The entire integration dies. I’m fairly experienced in troubleshooting device/network issues, controller issues, etc.
I’m a bit confused in the lack of logging. It’s almost as if I’m not looking in the right place (the place I’ve used countless times for ZWave related troubleshooting). If the integration is having a hard time initializing, you’d think there would be logs telling me why.
Thanks again, all. I’ll try to address the thoughts here:
I’m not convinced the controller is bad but it is a few years old so it’s not impossible.
I have it and a Zigbee controller plugged in to a powered USB hub and shared to the VM.
Mount location hasn’t changed according to the VM Manager
If it were any of the above, I would expect to seer mount/connection issues to the controller in the logs.
Fair point on ZWave network health. All nodes look healthy and communicating but I’ll keep an eye on it. It’s a small home. Most connections are direct to the controller.
VM has 4GB dedicated to it from a 96GB host. According to HomeAssistant, it’s using about 50%.
I’m waiting for a failure again in the hopes that I’m just missing something in the logs. I’m an experienced user and have referred to the logs successfully countless times for bad controllers, etc. So I think I know where to look but who knows, I suppose! haha!
Log Level is currently set to Info (should I go more verbose!?). And, since I’ve started this thread, there hasn’t been a failure. lol
This sounds like the problem. However, I’m just running the normal / internal version of the ZWave JS Integration. Not the “UI” version where I could put it behind a proxy.
The normal / internal version is now the “UI” version since 1.0.0. They have been combined. If you look at the App, you’ll see a button to open the UI.
BTW, you have to have both an “App” and “Integration”. The App is what talks to the Z-wave controller hardware. The integration talks to the App using Websockets. This has always been the case, just the old “internal” App had no UI. In contrast, the internal Zigbee integration (ZHA) talks directly to the coordinator hardware, no App needed.
Thanks again. Just dropping by for an update. Since creating this thread, I’ve had no failures. So odd. If this was a transient problem, I apologize for taking peoples’ time.
