Sudden change in how long messages remain in the postfix outgoing queue

HippoMan

6/24/24, 6:26 PM

I'm using postfix 3.5.8 under Debian 11.

Starting recently -- approximately 3-5 days ago -- outgoing messages suddenly started to take a lot longer than before to leave the outgoing message queue (as listed by "postqueue -p"). Whereas up until recently, messages would normally leave the outgoing queue within 5-15 seconds, it is now taking 1-2 minutes before messages leave that queue.

I have not changed any of the postfix settings. And here are what I believe to be the pertinent "main.cf" settings (as listed by "postconf -p"), and these have been set this way for many months. Note that these are only my guess as to which settings are pertinent, and of course, I realize that there could be other settings that I might be overlooking ...

default_destination_rate_delay = 0s
default_recipient_refill_delay = 5s
deliver_lock_delay = 1s
dnsblog_reply_delay = 0s
in_flow_delay = 1s
maximal_backoff_time = 4000s
maximal_queue_lifetime = 5d
minimal_backoff_time = 60s
qmqpd_error_delay = 1s
queue_run_delay = 60s

Given that this is output from "postconf -p", most of these values are postfix defaults. Only "queue_run_delay" and "minimal_backoff_time" are explicitly set in "main.cf".

Again, none of these parameters have changed in a long time, despite the fact that since 3-5 days ago, messages started remaining much longer in the outgoing queue than before.

I'm guessing that other settings -- including perhaps configuration of non-postfix-related utilities on my server -- might be causing these new outgoing delays, although I have not knowingly made any network-related changes on my server.

What factors could be causing these recent outgoing-queue changes?

Please forgive me if I have overlooked something in the postfix docs which might explain this.

And thank you very much in advance.

0 + 2

postfix

queue

anx

7/4/24, 1:56 AM

Unless you have made unusual use of [stress-adaptive](https://www.postfix.org/STRESS_README.html) configuration, the one thing that would greatly impede swift deliveries is waiting for timeouts where immediate hard errors would be appropriate - milters, policy daemons, dns, that sort of thing. Having logs and configuration (see the [tag description](https://serverfault.com/tags/postfix/info)) might make this answerable.

HippoMan

7/5/24, 2:21 AM

Thank you very much. I posted configuration already. This configuration listing is in my original question. And there are no messages in my logs which relate to this, and so I didn't originally post any log messages. And anyway, the problem seems to have disappeared on its own a few days ago ... again, with no configuration changes at all. And in any case, my older log files from the time when this problem was occurring have now been overwritten, so I can't post them, anyway.

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: Sudden change in how long messages remain in the postfix outgoing queue

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.