listmonk: outgoing Message-ID uses localhost.localdomain

[filter]

Hi,

All campaign emails sent from the listmonk app leave with a
Message-ID header pointing at localhost.localdomain, e.g.:

  Message-ID: <1776433035469678859.1.4025817209405639301@localhost.localdomain>

This is a minor but real deliverability smell — some spam filters
have heuristics against it, and it's unambiguously a misconfiguration.

Root cause

listmonk uses the knadh/smtppool library. In
smtppool/email.go :: generateMessageID():

  h, err := os.Hostname()
  if err != nil || !strings.Contains(h, ".") {
      h = defaultHostname   // = "localhost.localdomain"
  }

https://github.com/knadh/smtppool/blob/master/email.go

app UUID (e.g. 9983cd4c-e53c-436e-af86-84bdbc749178) — no dot,
so smtppool falls back to the hardcoded localhost.localdomain.

The "EHLO hostname" setting in listmonk admin does not help —
smtppool uses it only for the SMTP handshake, not for Message-ID
generation (two separate fields in the library).

Proposed fix (Cloudron package, one-liner)

Set the container hostname to CLOUDRON_APP_DOMAIN — either via
--hostname at run time, or hostname "$CLOUDRON_APP_DOMAIN" at
the top of the start script. Message-ID then becomes
<…@listen.example.com> instead of @localhost.localdomain.

An upstream config option in listmonk/smtppool would be the proper
fix, but the Cloudron-side workaround resolves it immediately for
all Cloudron listmonk users.

Environment

• Cloudron: 9.0.0
• listmonk app image: cloudron/app.listmonk.cloudronapp:202603300219460000

Happy to test a patched build. Thanks!

[girish]

@filter great report. I see the bug even in our current listmonk campaign

I have to investigate what the best way to fix this is. The container name cannot be set to fqdn . This will result in many issues including an app having trouble reaching itself using https://fqdn (because now fqdn will resolve to container IP and bypass nginx and thus have no TLS).