v3.9.0 – Resilient state handling
Improvement
Resilient state handling: Entities now keep their last known state for up to 3 consecutive poll failures before becoming unavailable.
How it works
- Each successful poll resets the failure counter to 0
- On failure 1 and 2: previous state is kept, entities remain available
- On failure 3:
UpdateFailedis raised, entities become unavailable in HA - Log messages now include the failure count (e.g.
Connection lost to MTX (failure 2/3), keeping previous state)
This applies to both MTX and XMP44 coordinators and prevents brief network hiccups or slow device responses from flickering entity states.