The CHECK SYSTEM unique name is expected to be resolvable

[Former user]

According to the documentation, the CHECK SYSTEM unique name may be anything: "The unique name is usually the local host name, but any descriptive name can be used."

However if it's not resolvable, monit will log errors when alerting: Jun 20 10:12:35 sandbox2 monit[5736]: Cannot translate 'foobar' to FQDN name -- Name or service not known

That error is raised from alert.c, _getFQDNhostname.

I suppose that either the documentation should be updated to reflect that the unique name MUST be resolvable, or the log error should be removed.

[Tildeslash]

Fixed already as part of #356.

[Cyril B.]

Sorry but it's not fixed by that commit, which is included in Monit 5.18 anyway (which I used).

[Tildeslash]

I'm sorry, yes, this warning will still be logged during alert. In practice it is very often error, as that name becomes part of alert sender address (by default "monit@$HOST") - if the domain doesn't resolve, many MTAs will refuse to deliver such email.

To set the correct sender address and get rid of the message, please set the mail-format:

set mail-format { from: [e-mail] }

[Cyril B.]

Actually I already do use a custom mail-format, and the error is still printed.

[Tildeslash]

what's the value of "from:" in your mail format? Please also make sure monit was reloaded when configuration changed (you can check the current runtime settings via monit HTTP GUI.

[Cyril B.]

Yes, monit was restarted. Relevant config:

#!
set mail-format {
   from: [e-mail]
   subject: [monit] (sandbox2.paris1) $SERVICE $EVENT
   message: $DATE on sandbox2.paris1
$ACTION $SERVICE: $DESCRIPTION.
 }
check system sandbox2foo.paris1

[Tildeslash]

found the problem, will fix

[Tildeslash]

Fixed: Issue #403: If the sender address for monit alert was overriden by "set mail-format" statement, don't try to resolve the hostname to FQDN.

→ <<cset 7508ddfa13be>>

[r1zar.it]

The same issue on current version 5.25.1 linux-x64

The conf is

set mail-format { from: [e-mail] }

check system mynotresolvingname
    if cpu usage (user) > 95% for 5 cycles then alert
    if loadavg (15min) > 15  then alert
    if memory usage > 90% for 5 cycles then alert
    group server-resources

[Tildeslash]

@vblazhnov please can you provide more details (error output)?