I also have a pile of Home Seer switches and have been seeing this same pattern in my ZWave logs for months. The result has been very unreliable performance, if you try and perform some action while the network is being flooded it just isn’t going to happen. Like others here, this started sometime within the last 6 months or so. Prior to that, everything was stable and I’ve not added or moved any hardware, which only leaves the software as a possibility.
I dug into several possibility to correct this.
First, re-locating my Z-Wave radio didn’t seem to help any. I wrote this up a while back and thought maybe that was the culprit, so I temporarily used a USB port from the back of one of my VM hosts with no benefit. Back to the Z-Wave radio near the middle of the house over IP because that seemed better.
I ran a “Heal Network”, no dice.
Finally, I noticed I had enabled polling on the HS devices so I went and re-configured all HS switches to not be polled and then… everything works like it used to. Instant Status still sends me status updates so I don’t really need polling on these devices anyway. This isn’t 100% bulletproof, sometimes local dimming at the switch doesn’t show up in the HA UI, but on/off still is picked up by HA reliably and basically all my automations rely on scenes so it kinda doesn’t matter if HA knows the current dimmer value.
Not sure if this helps anyone, I don’t have pockets deep enough to go and pull all my switches out to fix a software problem so I had to come up with something cheaper 