Re: [monit] Fine Tunning a monit configuration

[Martin Pala]

You can run monit in verbose mode and verify what happened:

  monit -vI

The described problem could be startup synchronization bug in monit 
which was solved in upcoming monit-5.0.

You can get monit-5.0_beta7 here:
http://www.mmonit.com/monit/download/

To create binary RPM from source distribution it should be sufficient to 
download it and run:

  rpmbuild -tb monit 5.0_beta7.tar.gz



Martin



Andres Tarallo wrote:
> I have a bunch of Centos 5.2 Servers running Apache, I've installed 
> monit 4.9 (RPMs from DAG repository). These servers are heavily loaded 
> most of the day (average 1 min over 20, many hours a day). I keep 
> getting the following messages in my Mailbox:
>
> ** Subject httpd Timeout - httpd unmonitor on XXXXX: 'httpd' service 
> timed out and will not be checked anymore.
> ** Subject httpd Connection failed - httpd restart on XXXXXX: 'httpd' 
> failed protocol test [HTTP] at INET[WW.WW.WWW.ZZZ:80] via TCP.
> ** httpd Does not exist - httpd restart on XXXXXX: 'httpd' process is 
> not running.
>
> The last one really puzzles me, because Apache is actually running !!!!!
>
> My monit configuration file
>
> set daemon  180
> set logfile syslog facility log_daemon
> set mailserver mail.company.net <http://mail.company.net>
> set mail-format { from: address@hidden <mailto:address@hidden>
> subject: $SERVICE $EVENT
> message: $SERVICE $ACTION on $HOST: $DESCRIPTION.}
> set httpd port 2812 and
>      use address localhost  # only accept connection from localhost
>      allow localhost        # allow localhost to connect to the server and
> check system XXXXXX
>     if loadavg (1min) > 50 then alert
>     if loadavg (5min) > 75 then alert
>     if memory usage > 90% then alert
>     if cpu usage (user) > 99% then alert
>     if cpu usage (system) > 99% then alert
>     if cpu usage (wait) > 99% then alert
>     alert address@hidden <mailto:address@hidden>
>
> check file apache_bin with path /usr/sbin/httpd
>    group apache
>    if failed checksum then unmonitor
>    if failed permission 755 then unmonitor
>    if failed uid root then unmonitor
>    if failed gid root then unmonitor
>
> check process httpd with pidfile /var/run/httpd.pid
>     start program = "/etc/init.d/httpd start"  as uid 0 as gid 0
>     stop program  = "/etc/init.d/httpd stop" as uid 0 as gid 0
>     if cpu > 99% for 5 cycles then restart
>     if loadavg(5min) greater than 45 for 3 cycles then restart
>     if failed host WW.WW.WWW.ZZZ port 80 protocol HTTP request 
> "/site/page.php" timeout 15 seconds 10 cycles then restart
>     if 6 restarts within 10 cycles then timeout
>     alert address@hidden <mailto:address@hidden>
>     depends on apache_bin
>     group apache
>
> I'm pulling my hair, It doesn't work flawlesly. I receive many alerts, 
> even when the servers are working. Thanks
>
>
> ------------------------------------------------------------------------
>
> --
> To unsubscribe:
> http://lists.nongnu.org/mailman/listinfo/monit-general