Here’s the FAQ on fixing zwave repair error messages if that’s of any interest:
And, yes, I build the backbone first because that way the battery operated devices will be able to find all the available repeaters when they build their own address tables. Sometimes the best repeater is one floor up, for example. But if you want to go room by room and then do the repair utility at the end, you can.
One thing to note – – there is a huge difference between Z wave plus and earlier zwave generations. Zwave plus is much better at pairing in place. Even so, the device doesn’t know that a repeater exists unless either that repeater was on the network before the battery operated device was added or until you’ve done a repair.
As far as Max devices, the Z wave standard itself limited network to 232 nodes including the hub.
But there have been various smartthings – specific issues in the past at both 100 devices and 200 devices, but some of those applied to a specific device class. I honestly don’t know what the current status is on those issues.
My personal rule of thumb is that if you run the zwave repair and you get errors on 10% or fewer of the devices, then fix one device at a time.
More than that, and I start looking at it as a damaged network. Typically I would first remove any devices added in the last week. Then take the hub off power for 15 minutes and then restart it, just to get it synced up again with the cloud account. Then then run a repair and see if that got rid of the problems.
Otherwise, I would look for a failing repeater than a number of other devices were trying to use.
But if you’re above that 10% threshold it just gets to be trial and error no matter which thread in the tangle you start pulling on first.