Snoopy
Member
Running Home Assistant on official Green hardware, integrated with my Kaleidescape (Strato C). A few times a day my HA instance goes completely unreachable (disconnect and reboot) for ~5 minutes, then recovers on its own — no manual reboot needed. It seems to happen especially right after pushing buttons on my dashboard related to KScape and Claude has narrowed it down to the Kaleidescape integration. I’m not sure if that is accurate or not.
Digging through the Supervisor logs, I found the actual cause: the watchdog misses two Core API responses in a row and force-restarts Core. That correlates directly with repeated "blocking call" warnings pointing at the Kaleidescape integration's device.connect() — it looks like it's doing synchronous file reads inside the async event loop, which stalls Core long enough to miss the watchdog pings.
Has anyone else running the Kaleidescape integration with a Strato C seen this? Curious whether:
1. Others get the same "blocking call" warnings from device.connect()
2. There's a known fix or workaround (different polling interval, alternate integration/fork, network setup change, etc.)
3. This is specific to Strato C or shows up on other Kaleidescape models too
Happy to share full log snippets if that helps narrow it down. I’m not very smart on these things, so it maybe not be Kaleidescape doing it. Thanks!
Digging through the Supervisor logs, I found the actual cause: the watchdog misses two Core API responses in a row and force-restarts Core. That correlates directly with repeated "blocking call" warnings pointing at the Kaleidescape integration's device.connect() — it looks like it's doing synchronous file reads inside the async event loop, which stalls Core long enough to miss the watchdog pings.
Has anyone else running the Kaleidescape integration with a Strato C seen this? Curious whether:
1. Others get the same "blocking call" warnings from device.connect()
2. There's a known fix or workaround (different polling interval, alternate integration/fork, network setup change, etc.)
3. This is specific to Strato C or shows up on other Kaleidescape models too
Happy to share full log snippets if that helps narrow it down. I’m not very smart on these things, so it maybe not be Kaleidescape doing it. Thanks!
