Hey folks,
I’ve recently set up Home Assistant Supervised on generic Linux with DynDNS + SSL. I’ve got a 443->8123 port forward on my router, and it all Just Works which is pretty neat! I’m running this on a media server (pretty overpowered vs recommended specs) which has a few other things going on (Plex and friends) in Docker compose containers. All of them are rock solid in terms of reliability.
I’ve noticed when I try to access HA via https://my.domain, it’s a bit unreliable. Often it loads fine; other times I get bounced to /lovelace with a loading spinner, then it times out. Back in my day this would have indicated connectivity, but I don’t know how HTML5 and offline containers and stuff work - maybe it can’t reach the server at all and it loads the spinner etc from cache?
The Chrome console gives me this:
Failed to load resource: net::ERR_CONNECTION_TIMED_OUT
manifest.json:1 Failed to load resource: net::ERR_FAILED
core.abbXXXXX.js:1 WebSocket connection to 'wss://my.domain/api/websocket' failed: Error in connection establishment: net::ERR_CONNECTION_TIMED_OUT
e @ core.abbXXXXX.js:1
lovelace:1 Uncaught (in promise) 1
service_worker.js:1 Uncaught (in promise) TypeError: Failed to fetch
How can I debug the issue here - for starters, I’d just like to confirm if the server is reachable but timing out on a HA response, or if it’s actually not even reachable. Beyond that - are there any common issues in this space? I’m logged in on a few devices - is there a cap somewhere I need to increase? As an interesting observation, it won’t respond to ping during the up or down periods.
I can see in home_assistant.log that various chromecasts and smart speakers periodically have heartbeat timeouts, but nothing that looks alarming or related to the web service.
Thanks in advance! Lovely bit of software.