I’ve only seen one report in the forums from someone who said that after they did a heal none of the zigbee devices responded. But this was at a time when a lot of people reporting problems with zigbee devices, and I believe the person did the heal because they were having problems already. So I can’t say for sure, obviously, but I don’t think the heal was the sole contributing factor.
Best practices per the independent standard are to update the neighbor tables as part of regular maintenance. For example, some of SmartThings’ competitors do an automatic zwave repair every night. Others recommend one once a week. It’s just a way of making sure that all the devices are functioning optimally.
That said, at one point in the summer of 2016 some SmartThings support staff were telling people that a zwave repair could make things worse in some situations. There’s absolutely no way that should be true for a certified Z wave controller. It should always be a “can’t hurt, might help” kind of thing. So I can only assume that there’s something strange about the cloud-based platform architecture that can lead to this conclusion.
I will note that since the issue was first raised, support no longer seems to say that. (Which is good, because it was giving me a headache. ) but again, once you introduce the cloud and various sync issues, who knows?
So I can’t tell you what’s the best thing to do for a SmartThings account. You can check with support if you have any concerns. General best practices for a Z wave or zigbee system would be to include these updates as part of periodic maintenance.
One thing I can say is that if you add a new Z wave repeating device to your existing network and you don’t do a zwave repair a lot of the other devices are not going to know it exists. So that’s not going to help your network efficiency very much.