Podman logs and priority level

[skoobasteeve]

Hello,
I'd like it if someone could help me understand podman's role in determining log priority levels when using --log-driver=journald combined with a syslog server. For some reason, 99% of the logs from my three containers (nextcloud-fpm, caddy, mariadb) are interpreted as "error" level by rsyslog. See attached examples as they appear on my LibreNMS server:

These appear to be standard info logs from the containers. How is rsyslog deciding that these logs should be error level? Not sure if I should be pursuing this with podman, rsyslog, or at the container level. Any guidance would be appreciated.

Thank you!

[Luap99]

The log driver only sees stdout and stderr from the container process.
stdout is logged at info level while stderr is logged at error level, see
https://github.com/containers/conmon/blob/2832e4f751939b4c63e34b2b6bcfbc103b1c7321/src/ctr_logging.c#L255-L258

[skoobasteeve]

So if I'm seeing these come in as err does that mean that the container is choosing so send this output to stderr instead of stdout?

[Luap99]

yes

[skoobasteeve]

Great, thanks for helping me understand. Is there a technical reason why a container author would choose to send seemingly routine messages to stderr instead of stdout?