heya 
using incus to manage some service on my personal server, loving it.
I’ve been having a bit of a dns issue with some containers – so I’ve got a traefik app container (linking to the dockerhub image) in a managed network with all of my other containers. Traefik serves as the entrypoint for port 80/443, then dispatches stuff to the other containers through dns (e.g. http://foo.lxd). Works well enough.
Occasionally though, some container or other will have its dns disappear from traefik’s sight, for unexplainable reasons… this was the case as well before when I used haproxy in an ubuntu system container. Trying a ping just gives a hostname not found error, even though the container is still visible if I pull an “incus network show lxdbr0”. Other containers are fine when this happens. Restarting the container usually fixes this. (the one this is happening to right now is an nginx app container.)
I’m using dns.mode=managed and no other particular config on an ubuntu host OS – I was wondering if there was anything about this that would be amiss, or if there was a way to get more details on what could be going wrong? I’m not finding any leads at the moment since everything seems “right” from the host OS side.
All my containers have the host machine pointed to as a dns provider in their resolv.conf, it seems (in this case 10.2.162.1 for an lxdbr0 network whose ipv4.address = 10.2.162.1/24).
Thanks in advance!