Dropped Z-Wave events in large networks?

I have finally decided to tackle an issue of occasional dropped Z-Wave events that my Hubitat setup has had since the beginning---I used to think an occasional dropped Z-Wave event is inevitable, but someone's response on another thread ((Re)set State of Software Sensors) suggests that I should perhaps expect (and work to reach) 100% reliability of Z-Wave event transmissions to the hub.

So I guess I would ask members of the community who have large Z-Wave networks (on the order of 100 devices or more): Are you able to achieve 100% reliability with Z-Wave event transmission in a large network? If so, have you had to troubleshoot issues of dropped Z-Wave events in the past, and what steps did you take to resolve the issues?

My background/context is that I have a mesh network with 120+ Z-Wave devices, almost all of which are Z-Wave Plus and about half of which are hardwired repeaters, in a two-story 2000-square-foot house (so the mesh network has a lot of redundancy to cover not a particularly large space). I don't have any other Z-Wave hubs or devices or any baby monitor or cordless phone causing interference, and I'm not aware of any neighbors with Z-Wave devices, baby monitors, or cordless phones. As a general practice, I repair the Z-Wave network after each add/remove of a repeater.

Nevertheless, I occasionally have some sensors of different types and brands (all Z-Wave Plus, all listed by the Hubitat website as officially supported devices for the platform, all with newish batteries reported at >95%, and all running the system default drivers) that get out of sync with Hubitat. For example, last night I installed a new Ecolink open/close sensor on a window about 5 feet above my hub, and it is currently out of sync: The magnetic contact is closed, but Hubitat says it's open.

I suspect that this discrepancy and others like it arise because a small number of Z-Wave transmissions occasionally fail to reach the hub, causing a sensor to appear "stuck" in a particular state in Hubitat. Is this normal?