Integrations don't connect to internet anymore

Configuration
1

Since a couple a days (I think since I updated to 2024.10), integrations which need to retrieve data via an API, show an error that no connection can be made (http 400). When I restart HA, initially it works, but after a few minutes, and after reloading an integration, the connection errors appear again.

At the end of this post. I put a few examples of errors in the log file.

I have an external connection to HA with Duckdns and Nginx Proxy Manager. I enabled Websockets support in Nginx.

afbeelding afbeelding

Other info which might be relevant:

  • Home Assistant Yellow
    • Core 2024.10.2
    • Supervisor 2024.10.0
    • Operating System 13.1
    • Frontend 20241002.3
  • I run Adguard Home (ad blocker) in my HA installation, but disabling AGH doens’t solve the problem.
  • Router is a Fritz!Box 7590.
  • In configuration.yaml:
    http:
use_x_forwarded_for: true
trusted_proxies:
  - 172.30.33.0/24
  - 192.168.1.49
  - 192.168.1.36
ip_ban_enabled: true
login_attempts_threshold: 5

Some examples of error messages:

Samsung smart things
afbeelding

2024-10-13 21:07:10.063 DEBUG (MainThread) [homeassistant.components.smartthings] 400, message="Expected HTTP/:\n\n  b''\n    ^", url='https://api.smartthings.com/v1/installedapps/.....'

Brink ventilation custom component

2024-10-13 21:12:37.588 DEBUG (MainThread) [custom_components.brink_ventilation.core.brink_home_cloud] POST request: https://www.brink-home.com/portal/api/portal/UserLogon, data {'UserName': '******', 'Password': '*******'}
2024-10-13 21:12:37.674 ERROR (MainThread) [custom_components.brink_ventilation.core.brink_home_cloud] Client error on API https://www.brink-home.com/portal/api/portal/UserLogon request Cannot write to closing transport
2024-10-13 21:12:37.675 ERROR (MainThread) [homeassistant.config_entries] Error setting up entry ***@***.** for brink_ventilation

Reolink doorbell

2024-10-13 21:10:42.280 ERROR (MainThread) [homeassistant.components.reolink.host] Reolink Deurbel event long polling subscription lost: Host 192.168.1.88:443: failed to subscribe long_poll: Host 192.168.1.88:443: subscription request got a response with wrong HTTP status 400: Bad Request

Buienradar (weather)

2024-10-13 21:10:56.563 WARNING (MainThread) [homeassistant.components.buienradar.util] Unable to retrieve json data from Buienradar (Msg: 400, message="Expected HTTP/:\n\n  b''\n    ^", url='https://data.buienradar.nl/2.0/feed/json', status: None)
2024-10-13 21:10:56.569 WARNING (MainThread) [homeassistant.components.buienradar.util] Unable to retrieve json data from Buienradar (Msg: 400, message="Expected HTTP/:\n\n  b''\n    ^", url='https://data.buienradar.nl/2.0/feed/json', status: None)

Solardedge (solar panels)
afbeelding

2024-10-13 21:11:52.054 ERROR (MainThread) [homeassistant.components.solaredge] Could not retrieve details from SolarEdge API

Frank (energy provider)

2024-10-13 21:11:56.966 ERROR (MainThread) [custom_components.frank_energie.coordinator] Unexpected error fetching Frank Energie coordinator data
2

Hi m4th31st,

What you describe is often the symptoms of a messed up DNS server, where HA is trying to find IP addresses and can’t.
I suspect adguard. I know you said you disabled it, but things don’t work that way. You somewhere have things set to talk to adguard and it talks to the world. If adguard is down, or adguard is missing, or adguard is badly configured you will get the same result. Can’t find external websites.

Look closer there, I bet you will find the problem.

3

Thats strange, because an http 400 response means your request has hit the server but its replied saying the request is invalid. This might suggest its not adguard.

4

The log of the AdGuard addon is constantly throwing these errors, which I don’t really understand…

2024/10/14 17:25:30.558953 ERROR response received addr=172.30.32.3:53 proto=udp status="exchanging with 172.30.32.3:53 over udp: read udp 172.30.32.1:48516->172.30.32.3:53: i/o timeout"
2024/10/14 17:25:30.562374 [error] dnsproxy: exchange failed upstream=172.30.32.3:53 question=";63.1.168.192.in-addr.arpa.\tIN\t SOA" duration=2.005126709s err="exchanging with 172.30.32.3:53 over udp: read udp 172.30.32.1:48516->172.30.32.3:53: i/o timeout"

This is the DNS configuration in Adguard:

afbeelding
afbeelding1160×612 13.8 KB

afbeelding
afbeelding1139×844 21 KB

Should I change something there?

5

After reinstalling AdGuard, I noticed there’s no checkmark at “Use Private reverse DNS resolver”. I don’t really understand what it does, but since “SOA” and “ARPA” also appear in the error messages, I think this was the cause of all trouble. And so far, all integrations are still behaving…
Anyway, thanks for putting me on the right track, @Sir_Goodenough

6

That the problem with adguard and pihole.
That’s why I do not suggest anyone use it unless you understand what it does.

It provides a man-in-the-middle attack on your personal data to supposedly “help” you, but always does what people tell it and people have no idea what they are telling it to do.

Result is this cry for help repeated over and over.

7

I understand what PiHole/Adguard does. I was talking about this specific option (private reserve DNS resolver) when I wrote that I don’t understand what it is used for.
It seems like you’re saying that someone has to understand every technical detail of a system to be allowed to use it.