I doubt there’s anything else we can do then as the binding shouldn’t be polling now (and it’s been 2 weeks with no problems, so I expect the binding is ok). Ultimately there’s a problem with the device, so if something manages to send a request it apparently will respond incorrectly.
As the timeout / false positive was happening during automatic scheduled wakeup phase (every 86400s/24h), I solved the issue by setting “Wakeup Interval” to “0” (never wakeup automatically).
No more false positives here after reset en re-inclusion. Thanks!