anarcat and I verified that the mails are coming in to eugeni properly, and getting forwarded to gettor@rdsys-frontend properly, and delivered to an inbox on rdsys-frontend properly.
Past that, we're not really sure what the service is supposed to do with it.
We looked around for a survival guide for rdsys-frontend but did not find one. :(
1059 messages in 2 days, in a fast look in prometheus I think 2/3 of those produced a help email and 1/3 did provide proper response with links. So maybe a big part of those are actually spam.
I can't see any clear bug introduced in the recent code or any obvious bug on the email processing code. I'm adding some more log output and I'll be watching the logs to see if hapens again.
I updated the imap library we use, just in case was a bug there (!54 (merged)). They've included in the main library the extensions we were using to idle on the imap server.
When you restart it, does it automatically go back and handle the mails it hadn't yet handled? That is, did the outage result in dropped replies, or just delayed replies?
I'm giving it 10 more days, if it doesn't fail again I'll close this issue and hope that is being fixed by the changes I've done. We'll hold the retirement of gettor-01 until then (tpo/tpa/team#40915 (closed))