AdGuardHome AdguardHome errors: dnsproxy timeouts?

[Bogey]

Recently these errors show up in my log:

Mar 16 09:29:44 RT-AX88U AdGuardHome[3233511]: 2024/03/16 09:29:44.034404 [error] dnsproxy: 127.0.0.1:53535: response received over udp: "exchanging with 127.0.0.1:53535 over udp: read udp 127.0.0.1:52186->127.0.0.1:53535: i/o timeout"
Mar 16 09:29:44 RT-AX88U AdGuardHome[3233511]: 2024/03/16 09:29:44.034574 [error] dnsproxy: 127.0.0.1:53535: response received over udp: "exchanging with 127.0.0.1:53535 over udp: read udp 127.0.0.1:45156->127.0.0.1:53535: i/o timeout"
Mar 16 09:29:44 RT-AX88U AdGuardHome[3233511]: 2024/03/16 09:29:44.034803 [error] dnsproxy: tcp://127.0.0.1:53535: response received over tcp: "exchanging with tcp://127.0.0.1:53535 over tcp: read tcp 127.0.0.1:59489->127.0.0.1:53535: i/o timeout"
Mar 16 09:29:44 RT-AX88U AdGuardHome[3233511]: 2024/03/16 09:29:44.034985 [error] dnsproxy: tcp://127.0.0.1:53535: response received over tcp: "exchanging with tcp://127.0.0.1:53535 over tcp: read tcp 127.0.0.1:59490->127.0.0.1:53535: i/o timeout"

I have Unbound running too, but on the surface everything seems fine.
How can I fix this?

 

[SomeWhereOverTheRainBow]

Recently these errors show up in my log:

Mar 16 09:29:44 RT-AX88U AdGuardHome[3233511]: 2024/03/16 09:29:44.034404 [error] dnsproxy: 127.0.0.1:53535: response received over udp: "exchanging with 127.0.0.1:53535 over udp: read udp 127.0.0.1:52186->127.0.0.1:53535: i/o timeout"
Mar 16 09:29:44 RT-AX88U AdGuardHome[3233511]: 2024/03/16 09:29:44.034574 [error] dnsproxy: 127.0.0.1:53535: response received over udp: "exchanging with 127.0.0.1:53535 over udp: read udp 127.0.0.1:45156->127.0.0.1:53535: i/o timeout"
Mar 16 09:29:44 RT-AX88U AdGuardHome[3233511]: 2024/03/16 09:29:44.034803 [error] dnsproxy: tcp://127.0.0.1:53535: response received over tcp: "exchanging with tcp://127.0.0.1:53535 over tcp: read tcp 127.0.0.1:59489->127.0.0.1:53535: i/o timeout"
Mar 16 09:29:44 RT-AX88U AdGuardHome[3233511]: 2024/03/16 09:29:44.034985 [error] dnsproxy: tcp://127.0.0.1:53535: response received over tcp: "exchanging with tcp://127.0.0.1:53535 over tcp: read tcp 127.0.0.1:59490->127.0.0.1:53535: i/o timeout"

I have Unbound running too, but on the surface everything seems fine.

I would venture to say this is more related to unbound taking too long to provide AdGuardHome with a response for a request. Either unbound needs a restart, or something is causing unbound to take longer than it should to provide an answer for a request. Possible reasons for unbound taking too long to respond could be related to firewall (e.g. skynet blocking access to certain regions request are being made to). That's just my first guess on why unbound could be taking too long. It is also possible unbound is not responding when certain domains are requested. Could be an issue with non-overlapping blocklists. An example being www.somedomain.com is not blocked by AdGuardHome, but it is blocked by Unbound; so, unbound is not responding to requests for www.somedomain.com (this is why I recommend disabling unbound-manager adblock features when using AdGuardHome to forward requests to unbound--- only adblock/filter requests with AdGuardHome). It could also be issues with overlapping dnssec options, or unbounds access control or privacy features are preventing the requests.

 

[SomeWhereOverTheRainBow]

How can I fix this?

In order for you to fix this, you need to determine what is causing unbound to not respond. Is it your firewall dropping request to IP addresses in blocked regions of the world? Is it conflict with unbound blocklists, privacy, access control, or dnssec options? It is going to require you to take a careful review of your firewall, and unbound query logs around the same time the request is being timedout in adguardhome. If it is not causing you DNS performance issues, I would not worry about it. Requests get dropped all the time. Most the time you are not even aware of the request timeouts or drops because these are "hidden" when it comes to standard logging practices. These would only be visible in "debug" or "error" logging for doing exactly what you are trying to do-- understand why a request is timing out.