Network issues - HA looses connection for few seconds

Installation Home Assistant OS
1

Hi. Recently I migrated my installation from RPi4 to Radxa X4. After some initial issues with migration (this was the first time I was doing something like that), I successfully launched it.
After few days of running it, I’ve noticed that sometimes it looses connection: WebUI has info about Reconnecting and if I was logged into CLI, it disconnects. At most cases hitting reconnect works but after some time it looses connection again. At this can last for few minutes or even much longer. I can’t find anything odd happening before that. The only two things I found in logs are:

2024-11-19 09:58:09.564 homeassistant NetworkManager[437]: <info>  [1732010289.5641] manager: NetworkManager state is now CONNECTED_SITE
2024-11-19 09:58:09.581 homeassistant systemd[1]: Starting Network Manager Script Dispatcher Service...
2024-11-19 09:58:09.594 homeassistant systemd[1]: Started Network Manager Script Dispatcher Service.
2024-11-19 09:58:14.140 homeassistant NetworkManager[437]: <info>  [1732010294.1403] manager: NetworkManager state is now CONNECTED_GLOBAL
2024-11-19 09:58:24.162 homeassistant systemd[1]: NetworkManager-dispatcher.service: Deactivated successfully.

But it’s not always there, so I don’t if it’s relevant.
When issue happens I get a lot (and I mean A LOT) of errors like this one:

0.32.3:48087" dns-server="udp:192.168.2.1:53" error="read udp 172.30.32.3:48087->192.168.2.1:53: i/o timeout" question=";api.telegram.org.\tIN\t AAAA"

Specific domain doesn’t matter because there are others, like api.spotify.com - I assume all DNS queries fails at this point. But Telegram one has the most as I see. Direct DNS log shows:

[ERROR] plugin/errors: 2 api.telegram.org. AAAA: read udp 127.0.0.1:59630->127.0.0.1:5553: i/o timeout
[INFO] 127.0.0.1:49596 - 55565 "NS IN . udp 17 false 512" REFUSED qr,aa,rd 17 0.00030491s

So far I’ve seen this issue three times and it calms down after a while. But that doesn’t mean it wasn’t there more often, I could simply miss it. Restart doesn’t help. No other device on the network has any issues at the time.

As for the config:
Network is set to static and currently uses Wi-Fi. I can’t check LAN for the moment.

System Information

version core-2024.11.2
installation_type Home Assistant OS
dev false
hassio true
docker true
user root
virtualenv false
python_version 3.12.4
os_name Linux
os_version 6.6.54-haos
arch x86_64
Home Assistant Cloud
logged_in false
can_reach_cert_server ok
can_reach_cloud_auth ok
can_reach_cloud ok
Home Assistant Supervisor
host_os Home Assistant OS 13.2
update_channel stable
supervisor_version supervisor-2024.11.2
agent_version 1.6.0
docker_version 27.2.0
disk_total 116.7 GB
disk_used 9.0 GB
healthy true
supported true
host_connectivity true
supervisor_connectivity true
ntp_synchronized true
virtualization
board generic-x86-64
supervisor_api ok
version_api ok
installed_addons CEC Scanner (3.0), File editor (5.8.0), Terminal & SSH (9.15.0), Mosquitto broker (6.4.1), Let’s Encrypt (5.2.3), etrv2mqtt (0.03), Piper (1.5.2), Whisper (2.2.0), Everything Presence Zone Configurator (1.1.0), Studio Code Server (5.17.3)
Dashboards
dashboards 2
resources 2
views 6
mode storage
Recorder
oldest_recorder_run 15 June 2023 at 17:08
current_recorder_run 19 November 2024 at 11:08
estimated_db_size 18173.64 MiB
database_engine mysql
database_version 11.6.2
Spotify
api_endpoint_reachable ok
2

Not sure if it’s related but I’ve seen several times this in the logs:

2024-11-20 18:53:35.185 homeassistant wpa_supplicant[452]: TDLS: Creating peer entry for xx:xx:xx:xx:xx:xx
2024-11-20 18:53:35.204 homeassistant wpa_supplicant[452]: TDLS: Dialog Token in TPK M1 2
2024-11-20 18:53:35.251 homeassistant wpa_supplicant[452]: nl80211: kernel reports: key addition failed
2024-11-20 18:53:35.251 homeassistant wpa_supplicant[452]: TDLS: Failed to set TPK to the driver

Where that MAC is from one of the phones that has HA installed. And it’s spamming logs for at least 2h straight, every . The internal IP was banned

3

Another “not sure if it’s related”: Because there were so many api.telegram.org entries, I decided to disable integration to see what that will bring. From that time I didn’t found that issue again. It could occur but at least it wasn’t prominent enough. I’ve seen once some DNS errors regarding Spotify’s API, I think. But there was no disruption to the net access. So I wonder if there’s some stuff happening when there’s too much DNS errors?

Anyway, since then I’ve manage to setup LAN. If it will be stable on it’s own I will re-enable Telegram to check if this issue reappear.