how to find this “chatty” device?
I do have an unknown node.
I am not able to remove that node somehow, not knowing where it comes from.
Can dat cause the problem?
Node002 - just look at the log you posted 
Look in Developer tools → States at your zwave. entities, and find the one that has an attribute of node_id: 2
I know the one 
I will reset it and add it again
A pitty. No succes again.
It looks like HA is not usable for me this way.
Zigbee and Zwave are the most important part of my domotica.
Very strange this happens. In Homey zwave worked fine. I had exactly the same setup in devices and flows
In day time everything is working oke.
In the evening there is some overload caused by something I don’t know and can’t find.
Maybe I give it a try to switch from Zwave dongle (AEONLABS).
and start all over again.
Is there a way to reset the zwave dongle?
I realy don’t know what the best is to do.
Any advice?
by the way: thanks for al effort you gave so far
Replacing dongles etc won’t help. Node 2 is flooding your network. Check config options of node 2 about reporting frequency or percent change etc you need to stop it getting so chatty.
1 Like
Thanks for the answer.
Yesterday I did unpair the node, resetted it and paired it again.
After that it was not chatty anymore.
But the network was still very slow and not reliable.
I will “tail” the messages of the controller for a while
In the mean time I did change all wallplugs the report by value change from 30% to 80%
and changed report time from 300 to 900 seconds
For the rest I don’t see interval consuming settings
At this moment the reports are coming in very slow.
Sometimes 10 sec. or there must be a PIR detection or so.
So at this moment the amount of traffic looks fine.
Every thing is reacting also as expected.
I will check again at night
Just a question:
I do have about 30 zwave devices. about half are router types (220V).
Is this much or just normal average amount of devices?
EDIT:
It is night and the most off my lights are running or activated by a PIR (Neo Coolcam).
The amount of traffic is still not realy high (I think it is about 1 message every 2 seconds or so)
Can it be something by the priority handling of the PIR’s?
Because the problem is that the lights are not switching on, on activating a PIR
And how to change that?
I do see this at the node settings at entity:
Do I need to change that setting?
EDIT 2:
I did change setting of all neo coolcam plugs to power change% 100% and update sequence to 1800sec. (I do have 10 Neo Coolcam wallplugs)
There is some improvement.
Sometimes the light is triggered within a second after PIR detection (1 out of 6 times or so) the other times it take up to 10 seconds to fire the light after trigger.
The improvement is that almost all actions are handled and non are lost.
The most device to trigger are Fibaro FGS223 switches (about 6) all are triggered by a Neo Coolcam PIR
The traffic is not that heavy in OZW-log. Sometimes every second a fire, sometimes 10 seconds nothing
I’m still not there but closer i feel 
It still is not giving the same results as with Homey.
The result is still not acceptable.
I just added the last device.
I do have 31 zwave devices now. It is to much for the zwave USB in combination with HA.
Starting after reboot takes about 25 minutes get all devices up and running.
I just ordered an Vera Edge.
I will use it as a Zwave hub and control it by HA.
I hope this will be te solution.
To be continued
I using now a VeraEdge as an extra Zwave Hub, beside my Zwave.me USB.
I did bring over 12 of the devices to Vera (out of 30).
I tried to keep 2 goed networkmeshes beside each other.
At this moment the Zwave is reliable and working fine.
I’m happy
I will do more experiences in the near future, to find out if I can understand why the unreliability is happening.
Now the situation is stable enough to go on