Hello Everyone,
diagnostics report: f299995c-83d0-40ba-b1d3-0698ae506947
All 3 of my ESPHome climate devices go Unavailable at the same time,
repeatedly, on v1.3.37. Every other Homey app keeps running, which is
why I think this is app-level rather than network or device.
Setup: 3x Midea AC with SLWF-01 Pro v2.1 (ESP8266), ESPHome 2026.7.2,
Noise encryption, no API password. Wired UniFi/MikroTik network, static
DHCP reservations, all three reachable on port 6053 from Homey during
the outages.
Measured with Homey Insights (uptime_sensor, 5-min resolution, 24h):
mideabianka 39 reboots / 24h (~every 41 min)
mideapanna uptime in minutes
mideaemelet-8c7896 0 reboots, 63h uptime ← control
The control device is on the same network, same firmware, and its
signal is only 2-5 dB better. WiFi is ruled out: only 4 of 39 reboots
were preceded by a >=5 dB signal drop, and over 7 days it was 0/20 -
reboots happened at signal levels 7-11 dB BETTER than the median. So
the device reboots look like a CONSEQUENCE of the app losing its API
connection (api reboot_timeout, 15 min default), not a cause.
Timing: the reboot rate jumped ~7x (1.5/day → 11/day) on Aug 11, the
day v1.3.34 was released. v1.3.37 (“overlapping reconnection attempts”)
looked like the right fix, but the devices dropped again after it.
Possibly related, possibly separate: the “host” setting of one climate
device keeps getting rewritten to the IP of an Everything Presence One
sensor that is NOT paired in this app. It does not hold a stale value -
it FOLLOWS that device: after I moved the EP1 from .95 to .89, the
climate device’s host became .89 within hours. Deleting the device and
re-pairing it did not help; the fresh record drifted the same day.
Meanwhile the other two devices keep their correct host.
Happy to run any test or provide more reports.