Outbound connection for Docker Container fail regularly

 Hi,

I have an issue with several docker containers that can not establish an outbound connection on a irregular, but frequent basis. 

To give you an idea how often this is, here a report for a high available service monitored with Uptime Kuma:

I'm also using Home Assistant as Container and receive error message like he one bewlow.

Logger: hass_nabucasa.iot
Source: runner.py:179
First occurred: 23. Mai 2023 um 08:25:37 (145 occurrences)
Last logged: 09:00:04

Unable to connect: Cannot connect to host cloud.nabucasa.com:443 ssl:default [Connect call failed ('3.65.231.73', 443)]

Cannot connect because unable to refresh token:

Unable to connect: Cannot connect to host cloud.nabucasa.com:443 ssl:default [Connect call failed ('3.68.57.87', 443)]

Unable to connect: Cannot connect to host cloud.nabucasa.com:443 ssl:default [Connect call failed ('35.158.69.194', 443)]

Unable to connect: Cannot connect to host cloud.nabucasa.com:443 ssl:default [Connect call failed ('3.64.223.38', 443)]

I've been thinking that it's maybe a DNS problem, witch is why I changed the DNS configuration of Unraid to seperat public DNS servers, but I get sometimes issues timeout messages.

Can you point me to a place where I can try to resolve it of find more logs please?

Here some weird behavior from the console:

root@NUK:~# docker exec -i 213a9fcfbe6b nslookup google.de 
;; communications error to 9.9.9.11#53: timed out
Server:         9.9.9.11
Address:        9.9.9.11#53

Non-authoritative answer:
Name:   google.de
Address: 172.217.18.3
Name:   google.de
Address: 2a00:1450:4001:80b::2003

root@NUK:~# docker exec -i 213a9fcfbe6b ping google.de
PING google.de (172.217.18.3): 56 data bytes
64 bytes from 172.217.18.3: seq=0 ttl=119 time=6.882 ms
64 bytes from 172.217.18.3: seq=1 ttl=119 time=5.817 ms
64 bytes from 172.217.18.3: seq=2 ttl=119 time=5.772 ms
64 bytes from 172.217.18.3: seq=3 ttl=119 time=6.593 ms

root@NUK:~# ping google.de
PING google.de (172.217.18.3) 56(84) bytes of data.
64 bytes from fra24s22-in-f3.1e100.net (172.217.18.3): icmp_seq=1 ttl=119 time=7.15 ms
64 bytes from fra15s28-in-f3.1e100.net (172.217.18.3): icmp_seq=2 ttl=119 time=5.02 ms
64 bytes from fra15s28-in-f3.1e100.net (172.217.18.3): icmp_seq=3 ttl=119 time=5.24 ms
64 bytes from fra02s19-in-f3.1e100.net (172.217.18.3): icmp_seq=4 ttl=119 time=5.53 ms

root@NUK:~# nslookup cloud.nabucasa.com
;; communications error to 9.9.9.11#53: timed out
;; communications error to 9.9.9.11#53: timed out
;; communications error to 9.9.9.11#53: timed out
;; communications error to 149.112.112.10#53: timed out
Server:         10.0.0.4
Address:        10.0.0.4#53

Non-authoritative answer:
Name:   cloud.nabucasa.com
Address: 52.28.212.129
Name:   cloud.nabucasa.com
Address: 18.194.43.31
;; communications error to 9.9.9.11#53: timed out
;; communications error to 9.9.9.11#53: timed out
;; communications error to 9.9.9.11#53: timed out
;; communications error to 149.112.112.10#53: timed out

Ping from container and host work. A few ours later not anymore. I now did another nslookup to the Home Assistant Cloud where I loose many times the connection and I can not resolve DNS requests. I have configured 3 DNS server. 1 and 2 are public, 3 is local (10.0.0.4) AdGuard. 

I read somewhere that I should not configure a container as DNS server for Unraid, so I keept it only if the externals are off. I'm stuck with my linux skills to figure out how to narrow down the issue any further. 

If I can't get this resolved in the next few days, I might need to reconsider not buying Unraid, since I'm still on a Trail license. 

GitEye-2.2.0-linux.x86_64.zip

Edited May 25, 20232 yr by o.goldich
More debug infos added