Failed to receive response from Home Assistant for request.

[CraigHaynie]

I am using this integration in HA and throughout the day all my cameras say the above message at the same time. I also have one camera set to live and it says "Camera Unavailable" when this happens. This is what the error on screen says:

{ "request": { "type": "media_source/browse_media", "media_content_id": "media-source://frigate/frigate/event-search/snapshots////GateCam2//", "id": 213 }, "response": { "code": "unknown_error", "message": "Unknown error" } }

If I refresh the browser (sometimes two or three times) it will go back to normal.

Any ideas what to check?

[dermotduffy]

This is almost certainly the underlying Frigate backend (not Home Assistant) dying/restarting or otherwise being unavailable. If you check your Home Assistant log you'll probably find an exception is being thrown in talking to Frigate.

As for the underlying cause, check your Frigate logs and see if you can figure out what is causing it to restart / be unavailable.

PS: If you're using the Frigate card, it could be https://github.com/dermotduffy/frigate-hass-card/issues/911 -- where the Frigate backend appears to have trouble answering in time in certain instances.

[CraigHaynie]

Thanks for the quick response! Right after I posted this I was looking at my “todo” list for today. Another item there was to disable the secondary Ethernet port on my Supermicro server. I see log entries all the time where my interface seems to be jumping around. I haven’t dug deep yet because I know I do not use the other port so I plan to disable in BIOS and see if that issue resolves. As soon as I saw that todo I wondered to myself if it could also have to do with this issue as well. I am heading to do that in 20 min and will see if it resolves anything.

I looked in HA logs just now and saw these …

Error fetching frigate data: 8:25:47 PM – (ERROR) Frigate (custom integration) - message first occurred at January 15, 2023 at 10:51:24 PM and shows up 342 times Error fetching information from http://10.1.10.8:5000/api/stats: Cannot connect to host 10.1.10.8:5000 ssl:default [Connect call failed ('10.1.10.8', 5000)] 8:25:47 PM – (ERROR) Frigate (custom integration) - message first occurred at January 15, 2023 at 10:51:24 PM and shows up 9308 times Error from stream worker: Error demuxing stream: [Errno 5] I/O error: 'rtmp://10.1.10.8/live/PackageBox' 8:25:43 PM – (ERROR) Stream - message first occurred at 8:02:08 PM and shows up 24 times [140142560914704] Error handling message: Unknown error (unknown_error) from 10.1.10.219 (Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:108.0) Gecko/20100101 Firefox/108.0) 8:24:57 PM – (ERROR) Home Assistant WebSocket API (custom integration) - message first occurred at 8:11:36 AM and shows up 263 times Error fetching information from http://10.1.10.8:5000/api/events?camera=GateCam&limit=50&has_snapshot=1&include_thumbnails=0: Cannot connect to host 10.1.10.8:5000 ssl:default [Connect call failed ('10.1.10.8', 5000)] 8:24:57 PM – (ERROR) Frigate (custom integration) - message first occurred at 8:11:36 AM and shows up 263 times Error fetching 10.1.10.134 data: Unable to connect to the device: 10.1.10.134: [Errno 111] Connect call failed ('10.1.10.134', 9999) 8:21:09 PM – (ERROR) TP-Link Kasa Smart - message first occurred at 11:51:38 AM and shows up 8 times Cannot open connection tcp://10.1.10.8:1935?tcp_nodelay=0 8:17:45 PM – (ERROR) components/stream/worker.py

Craig

[dermotduffy]

If you were having intermittent connectivity issues on your Frigate machine it would probably cause this, yes. Also I notice even your RTMP streams are having connectivity issues, so it's likely not related to the other issue I linked to and more a general connectivity problem between HA & Frigate.

[CraigHaynie]

I am headed to turn the other Ethernet port off now. As for other connectivity issues … both HA and Frigate are VM’s on a Proxmox box. Both are likely communicating with one another without leaving the box. But, this Ethernet port thing could be causing some issues.

If you have other ideas I am all ears … I am headed to disable those ports now!

Craig

[CraigHaynie]

As it turns out, the other Ethernet port can’t be turned of in the bios. I am not sure what the heck is going on. I was at the terminal so couldn’t copy the part of the log that I think is the offender. I’ll have to try again tomorrow to see if I can figure out what the heck is up.

Craig

Message ID: @.***>

[CraigHaynie]

I cleared the logs and capture a fresh group of entries in HA -> logs.

`Error fetching information from http://10.1.10.8:5000/api/stats: Cannot connect to host 10.1.10.8:5000 ssl:default [Connect call failed ('10.1.10.8', 5000)]

6:07:46 PM – (ERROR) Frigate (custom integration) - message first occurred at 6:07:41 PM and shows up 2 times [140191020380608] Error handling message: Unknown error (unknown_error) from 10.1.10.20 (Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/109.0.0.0 Safari/537.36)

6:07:41 PM – (ERROR) Home Assistant WebSocket API (custom integration) - message first occurred at 6:07:41 PM and shows up 4 times Error fetching information from http://10.1.10.8:5000/api/events?camera=GateCam&limit=50&has_snapshot=1&include_thumbnails=0: Cannot connect to host 10.1.10.8:5000 ssl:default [Connect call failed ('10.1.10.8', 5000)]

6:07:41 PM – (ERROR) Frigate (custom integration) - message first occurred at 6:07:41 PM and shows up 4 times Error fetching frigate data:

6:07:41 PM – (ERROR) Frigate (custom integration) Error from stream worker: Error demuxing stream: [Errno 5] I/O error: 'rtmp://10.1.10.8/live/PackageBox'

6:07:40 PM – (ERROR) Stream Unable to find referenced entities switch.outdoor_pool_1 or it is/they are currently not available

5:56:48 PM – (WARNING) helpers/service.py`

In Frigate I am seeing:

[2023-01-23 18:07:40] frigate.watchdog INFO : Detection appears to have stopped. Exiting frigate...

I ordered a new PCI USB card and cable. I am going to try that.

[CraigHaynie]

SUCCESS!!!

I had ordered a new PCI USB card for my server. One that had a power input. I got it installed last night and have not had a single issue since!! My inference speed with the other card had been about 24. The first thing I noticed last night was the speed is now a little over 8! I wonder if I was having an under power issue. I am marking this closed