Several devices (routers) are not connected (linked) to coordinator/other router on map, but are available and managed and they are connected to network in real. They are listed as Connected via [Generic Zigbee Coordinator (EZSP)] in device info.
Is there any way to fix/rebuild map or, at least export network topology information to analyse in csv or other format?
You might try a different map.
Generally, though, trying to analyze and manage Zigbee is a rabbit hole it’s hardly worth venturing into. If it works, it works.
I’ve seen that happen when the device is new to the network or the device/coordinator have recently been rebooted. For me those always resolve themselves. As was mentioned, trying to actively manage a Zigbee network is a bit of a rabbit hole, so as long as the devices work, I probably wouldn’t worry about it.
I think that generally means it’s a device that it knows about, but didn’t respond when you built that map at that moment, so a battery device, a device that responds different, etc.
Just remember, don’t chase the maps, they are a maze a twisty little passages all different.
One of devices (MOES 2-gang switch - Zigbee router) is in this state for months as for now. Second added yesterday, still in this state, also MOES 2-gang switch (router). Both are wall powered devices. Several more similar MOES 2-gang switches are normally linked in topology.
I have devices that don’t respond when the map query happens, and they work fine.
I have others that go battery dead or don’t respond and they are dead.
I have battery devices that respond if they happen to be listening.
Also just because some path looks on the map to have a better signal, it might not use it.
I’m convinced someone wrote the first map function just to show off how messy their mesh was, and that’s about all it is good for.
I fell into the rabbit hole when trying to fix the map in a working Zigbee network.
Besides the “visual” map mishaps my primray issue was that I had to reinstall/reset most but not all and not allways the same battery powered sensors after each and every HA core update. Not just to have them turn green in the map but to have them online at all.
I managed to get out by switching (from scratch not migrating) from ZHA to Z2M. Network is fine after every update and every reboot. Map since shows what one would expect incl. mesh links.
No aditional routers, no relocation, no RF interference. Just bits and bytes here.
In MY case (and nothing but) I found ZHA failed not only regarding the map but in bringing up the network itself after HA core updates.
Besides that Z2M fully supports all my devices incl. less common EFH thermostats and various cheap sensors where ZHA did so only for the routers.
I’m not sure that any of the maps have much value. I have tried several; they give slightly different results which in every case are a snapshot of something I can’t control anyway. Meanwhile, the network is perfectly healthy.
They’re fun and very pretty, but if you really want to know about network quality an entities card containing the LQI values for all your devices will do the trick, and you’ll be able to watch it change in real time - something which the maps actually make more difficult.
As for ZHA vs Z2M - they’re just integrations. The network is exactly the same. And you still have no control over it.
Edit:
Only just registered the phrasing of the OP. It will depend on the size of your network, but most routers will probably not be connected to the coordinator - they connect through other routers. Was this your concern?
No. They just not linked to anything - neither to coordinator, nor to the other router on map, hanging in the air - that is the concern. But, only on the standard map, because they are available, connected to the network, sending data and receiving commands. Also links are more correctly displayed on dan-danache’s Zigbee Map app. Something wrong with the standard Zigbee integration app’s map behavior.
Yes, I certainly found the custom map more informative. I did find, though, that it listed devices with low LQI (<100 ish) as unavailable, whereas they were working and had valid LQI entities.
As I say, the real diagnostic value of all the maps is debatable, and since Zigbee is self regulating there’s not much one can do about it.
