Dovecot log is full of Watchdog info

mooCH

I just had a look at my Dovecot logs, and it’s majority full of Watchdog information like this:

02.04.2024 11:53:08	info	managesieve-login: Disconnected: Connection closed (no auth attempts in 0 secs): user=<>, rip=172.22.1.14, lip=172.22.1.250
02.04.2024 11:53:08	info	imap-login: Disconnected: Aborted login by logging out (no auth attempts in 0 secs): user=<>, rip=172.22.1.14, lip=172.22.1.250
02.04.2024 11:53:08	info	imap-login: Disconnected: Aborted login by logging out (no auth attempts in 0 secs): user=<>, rip=172.22.1.14, lip=172.22.1.250, TLS, TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)
02.04.2024 11:53:08	info	lmtp(1761): Disconnect from 172.22.1.14: Logged out (state=MAIL FROM)
02.04.2024 11:53:08	info	lmtp(1761): Connect from 172.22.1.14

Is there a way to avoid logging information from the Watchdog container (IP 172.22.1.14)?
I only want to see connection related information from the Internet.

HighWay777

I have the same problem.
Tried to disable watchdog email notification, but no effect
Apr 10 10:20:33 ac036091fc05 dovecot: imap-login: Disconnected: Aborted login by logging out (no auth attempts in 0 secs): user=<>, rip=172.22.1.13, lip=172.22.1.250, TLS, TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)

Apr 10 10:20:33 ac036091fc05 dovecot: imap-login: Disconnected: Aborted login by logging out (no auth attempts in 0 secs): user=<>, rip=172.22.1.13, lip=172.22.1.250

Apr 10 10:20:33 ac036091fc05 dovecot: managesieve-login: Disconnected: Connection closed (no auth attempts in 0 secs): user=<>, rip=172.22.1.13, lip=172.22.1.250

esackbauer

Is that really a problem? I guess its only on single mailbox installation. Once you have more users, and have your mailcow for months reachable via internet, your watchdog log entries are the smallest part of the total log.

HighWay777

esackbauer
So, If I correctly understand this is the way how watchdog checking dovecot functionality . In my case it happening on every 20 seconds. It will be good if there a way to reduce this check frequency.
Its not about the log size, but I think this check is more “aggressive” than it needed 🙂
I am coming from a “physical” email server which was worked fine during the last 15 years , and I never noticed any dovecot crash during this time.
Thank you for your reply

Regards,

HighWay

DocFraggle

HighWay777 In my case it happening on every 20 seconds

I just checked the watchdog.sh script, it sleeps between 20 and 79 seconds after each check. So if it’s checking every 20 seconds your RNG is broken… 🙂

$(( ( RANDOM % 60 ) + 20 ))

But I have to agree, if you could set the value which is currently 20 via variable it wouldn’t hurt

mooCH

Yeah, absolutely agree with HighWay777. I too come from a postfix/dovecot installation on a pure metal machine and never ever had an issue with dovecot crashing over several years.

Checking if the system is up every 20 seconds is total heresy!
There should be a way to set up the time interval in Watchdog.

And if “your watchdog log entries are the smallest part of the total log” as esackbauer said, I cannot even imagine going through that log. Only if you forced me to do it, because my eyes would be sore!