Help! Reply by email is randomly delayed

Hi everyone,
We’ve set up reply by email on our discourse and I noticed weird behavior. It’s often delayed by anywhere from 12 to 24 hours.
I’ve tried to test it myself (using two users I own) and the reply was posted immediately - so I’m quite confused.

I initially believed it could be because of the user’s inbox (or rather outbox) but I noticed that late replies tend to arrive together, even though they were sent at different times.
I also looked into seeing if the replies that arrived late were from the same domain but they weren’t. Gmail, Yahoo, iCloud etc.

Is there something we can change on our end?
Please help! Thanks.

1 Like

It sounds like an intermittent issue with your mail service. You can compare the discourse sent mail logs with your mail service logs.


Thanks @pfaffman!
I checked that further. As an example, two reply by email were sent 20 hours ago, and one was sent 18 hours ago but all were pushed to the forum only 10 hours ago. When I checked the “receive” log on the settings it says it was received 10 hours ago, but next to the message I can clearly see that the reply was sent long before that.
I am using discourse mail-receiver that I set up on my server. I’m trying to understand why it is “holding” and posting those replies with such a huge delay.

Any thoughts? Or troubleshoot you suggest?
Many thanks again!

1 Like

Well, I think that you can

 docker logs -f mail-receiver

to watch the mail receiver logs. You can send mail to anyaddress@yourhostname and watch how it gets processed.


I just looked at it, and it seems like some messages get stuck in the queue for 10-12 hours roughly, some get pushed immediately. I can’t find any reasoning for that.
After quite a few hours they’re “released” from the queue.

I might have found a solution, I need to confirm this. For some reason, the account that was set in the mail-receiver was not an admin but just a moderator. I’m not sure why, and if so why some emails weren’t delayed. I’ll hopefully report back!


It’s possible that this status meant that ratelimits weren’t fully disabled, or certain messages were undeliverable, forcing extra retries and queuing? I don’t think it’s been tested with a non-admin account, several things could have gone wrong.


Possibly. So far what I did was to send myself a message, and reply. Three email were failed and went into the queue to try later, after making sure the user is admin, the fourth email came through immediately. I’m going to double check it in the next few days as users reply via email.


Just to keep you all posted - I believe this was the case.
Further more, I believe the issue was on our end because of SSO. The admin parameter from SSO returned false, so with every log in the user’s admin status would be revoked. It seems to be okay now though. Many thanks for your help! Hope this helps others too.