[v0.108.0-b.58] Lots of errors: Client.Timeout exceeded while awaiting headers

[timkgh]

Prerequisites

• [X] I have checked the Wiki and Discussions and found no answer

• [X] I have searched other issues and found no duplicates

• [X] I want to report a bug and not ask a question or ask for help

• [X] I have set up AdGuard Home correctly and configured clients to use it. (Use the Discussions for help with installing and configuring clients.)

Platform (OS and CPU architecture)

Linux, MIPS LE

Installation

GitHub releases or script from README

Setup

On a router, DHCP is handled by the router

AdGuard Home version

v0.108.0-b.58

Action

Lots of timeout errors that I believe started with

2024/10/12 06:34:00.268414 ERROR response received addr=https://dns.controld.com:443/... proto=udp status="requesting https://dns.controld.com:443/...: Get_0rtt \"https://dns.controld.com:443/...?dns=AAA...AAA\": context deadline exceeded (Client.Timeout exceeded while awaiting headers)"

The network is fine, ping times to controld are 7-8ms.

What should the upstream_timeout be set to? I changed it from the default 10s to 1s because 10s is way too much to wait for a DNS query, even 1s is a lot. Are these errors related to upstream_timeout and maybe 0-RTT issues in quic-go?

It seems to use HTTP/3, is there a way to force it to HTTP/2 only?

Expected result

very few errors

Actual result

lots of errors

Additional information and/or screenshots

No response

[timkgh]

This seems to be a problem for HTTP/3 (UDP). Could it be a problem with 0-RTT requests and quic-go?

I switched to DoT (tls://) and not seeing any timeout errors so far. If there is a way to force HTTP/2 (TCP), happy to test that.

[SukkaW]

This seems to be a problem for HTTP/3 (UDP). Could it be a problem with 0-RTT requests and quic-go?

I doubt that. See my issue #7357, even DoT is falling and timeout for me, so it might have something to do with TLS rather than HTTP/3 or UDP.