ShellyForHASS: [BUG] Shelly 2.5 entities unavailable

Environment

• ShellyForHASS release with the issue: 0.2.1
• Last working ShellyForHASS release (if known): -
• Home Assistant Core release with the issue: 2021.2.3
• Operating environment (Home Assistant/Supervised/Docker/venv): Supervised

Describe the bug

It happens that a shelly dimmer 2.5 (v.1.9.4) which is configured as a roller shutter, that a subset of the entities become unavailable. At the moment I have a device where the following entities are unavailable:

• cover.shellyswitch25_kueche
• sensor.shellyswitch25_kueche_current_consumption
• binary_sensor.shellyswitch25_kueche_1_switch
• sensor.shellyswitch25_kueche_1_total_consumption
• sensor.shellyswitch25_kueche_2_total_consumption
• binary_sensor.shellyswitch25_kueche_2_switch

The following entities are still available:

• sensor.shellyswitch25_kueche_device_temp
• binary_sensor.shellyswitch25_kueche_over_temp
• sensor.shellyswitch25_kueche_rssi
• sensor.shellyswitch25_kueche_uptime
• sensor.shellyswitch25_kueche_voltage

I renamed all the entities manually. It seems to me that the “general” entities for the device are available and the ones specific to an actor or sensor are unavailable, but I don’t know if that is a valid classification and if it has been the same entities that were unavailable in the past. For the cover entity I am certain that is has been affected when I previously encountered this bug.

The available entities are getting regular updated values. Restarting home assistant doesn’t seem to recover the unavailable ones. The only thing helping is to reboot the device. Then all entities will be available for a while until it happens again.

I encountered this bug with another device in the previous weeks. To find a solution, today, I did a factory reset on the device and deleted the ShellyForHass integration and readded it, then renamed the enitites to the old values (so that my automations can use it again). The previously affected device is fine at the moment (since it was a few hours ago that may change), but now the device listed above has the same symptoms.

I’ve found the following log message: 2021-03-07 17:19:17 ERROR (CoAP) [pyShelly] Timeout connecting to http://192.168.90.80/settings [shellyswitch25-73C35E.fritz.box] Since that timestamp the device is unavailable. I couldn’t find log entries for that ip after that entry.

The web ui of the shelly is available and I can also access /settings and /status.

Steps to Reproduce

Wait a few days until some device has entities that are unavailable.

Expected behavior

Obviously, all entities should be available.

Screenshots

Traceback/Error logs

Additional context

status.json.txt settings.json.txt