See Automate Z-Wave JS Re-Interview? which has been ongoing since June 2021.
You don’t need to re-interview nodes unless there are driver changes requiring it, or there was a problem in the initial interview.
Thanks. That has been my solution to resurrect nodes which repeatedly go 
I will try a simple ping to see how that fairs…
This thread links to my original issue from last year. After the latest firmware for the zwave 700series sticks came out, this reduced the issue by 99%. I do still occasionally get failed nodes on my outside plugs, but a re-interview is no longer required. A simple ping works. And now all devices have a ping button created automatically by zwavejs, so I just setup an automation on those devices so when the state switches to unavailable it presses the ping button.
1 Like
There’s a large thread with a subject in automating pings. A re-interview is just doing the same thing, with a huge expense, and a possibility of making things worse.
2 Likes
I just had one of my iBlinds not close tonight. Pinging the node was fine, but it was unresponsive to any other command to control it. I did a re-interview and it immediately started responding. I therefore am marking this as unsolved…
A re-interview is just a band-aid though, I would suggest trying to get the root of the problem addressed.
All I have is band-aids! I have 3 x V2.0, 2 x V2.1 and 1 V3.0 iBlind motors. It’s always 1 of the V2.0 (kitchen_blind_2) and rarely 1 other V2.0 (kitchen_blind_3).
The graph appears to be wrong or I am misunderstanding it.
Node 76 only has 1 return route direct to the controller.
The node is line-of-sight 18’ 3" from its antenna to the controller antenna.
The firmware is not user upgradeable on V2.x, only V3.0 and newer, so the only option is to replace with something different, or buy a new iBlind…
These blinds tilt via pull strings so options are limited.
I would suggest that this is an issue for node-zwave-js, not HA (and especially not under a feature request topic). Unfortunately, this will be very difficult to debug if you are using the official Z-Wave JS add-on. You will need to switch to Z-Wave JS UI so you can capture debug logs to files over time. I assume the node goes dead, and sometime later you notice and do the re-interview? If you are actually able to get debug logs of the entire process (node goes dead, no responding to ping, re-interview fixes it) through HA, then I would submit an issue and include those logs.
I am using this add-on. . 
and this integration. . . . .
I have excluded the node and re-added it and now the graph shows only a single connection to the controller…
Great! Then you can capture debug logs as files, and refer back 7 days. https://zwave-js.github.io/zwave-js-ui/#/troubleshooting/generating-logs?id=driver-logs
Of course, with the new v12 Z-Wave JS sticking “basic” entities on literally all my light switches, now I need to re-interview dozens of devices and I there doesn’t seem to be any way to do that in bulk. I’m sure many other users are in the same boat.
You have dozens of Z-Wave light switches? Lucky you! I only have 6 (not dozen)!