Intermittent outbound SMTP email channel fatal Error


  • Used Zammad version: 6.3.0
  • Used Zammad installation type: (source, package, docker-compose, …) Package
  • Operating system: Debian 12
  • Browser + version: Firefox, Chrome, Edge or Brave latest verions.

Expected behavior:

  • Using Active Directory user mailbox and password connecting to Exchange with no authentication required (for internal use only). Both inbound and outbound use the same AD account. Both should stay connected

Actual behavior:

  • Inbound connection has always stayed connected. Could not get the Outbound to connect so used a AD admin account to get it to work. Then trying again to use the same account as the Inbound after upgrading to 6.3.0 the Outbound connection worked. After ‘X’ minutes or hours the Outbound connection will fail. Going thought the Edit process and both in and out will connect again for a period of time before Outbound fails again.

{“healthy”:false,“message”:“Channel: Email::Account out Net::SMTPFatalError”,“issues”:[“Channel: Email::Account out Net::SMTPFatalError”],“actions”:,“token”:“pXF_CttPeH7hNjuu5WUAyclmveU5hiZjGOmRSwqenElVCbghRh_xuA”}

Steps to reproduce the behavior:

  • Edit connection - all connects, wait then Outbound fails with above error.

Do I find the mail logs with the Production.log? Any suggestions for troubleshooting. Yes I am looking at the Exchange end of clues as well. Thanks

SMTPFatalError -usually- indicates a faulty email address.
The easiest way to find that one is to grep for “FatalError” through the production log (that should be good enough). You should be able to find the error message line, followed by a traceback. A little over that traceback you should find the culprit causing it.

Zammads Email channels are self healing. This means that the next successful operation will clear the fault. Try to send a email over that channel - if that clears the error, it indicates it’s either something very temporary in the interconnect or a specific user causing it.

Thanks for the reply… It’s strange, it doesn’t repair but will reset only if I go through the ‘Edit’ account process. Then at some point it will fail again.
The channel is using the same account for inbound as it is for outbound. It is only the Outbound that fails. I’ve switched the Outbound to use my personal credentials and that works just fine. Except you can’t send a email to my account and have it generate and open new tickets like it does with the support@ account. That might be a setting I am forgetting right off the top of my head.

Are we really talking about a regular communication channel and not a notification channel? Are you really sure that you are using the correct ticket group to send via the channel in question? Because if the sendout is successful (which might take a minute or two… remember… delayed action).

If all that doesn’t help… production.log.

I’ll do a deep dive tomorrow with your suggestions and report back. Thanks again.