Once in a while a mysqld running in this environment will just go a little
nuts and use up 50-100% of an entire processor.

It's not a big deal, but you probably are not getting peak performance out
of your mysql server when it is in that state.  Luckily, simply restarting
it will solve the problem - could you please do that ASAP ?

We have put together a little script that will do this for you
automatically.  We would appreciate it greatly if you could implement it:

First, put this line of code into a script /root/mysql.check

top -d10 -s12 | grep mysqld | awk -f /root/check.awk

then save the file and chmod +x it.  Then add this line to your root
crontab, by running, as root:  crontab -e  and putting this line into
the crontab:

*/15 * * * * /root/mysql.check

Then, create the file /root/check.awk and put this in it:

BEGIN {
        count = 0
        naptime = 5
}

{
        if($10 > 60)
        {
                ++count
        }

        report = sprintf("%s\n%s",report,$0)
}

END {
        if(count > 8)
        {
                system("/usr/local/etc/rc.d/mysql-server.sh stop")
                system(sprintf("sleep %d", naptime))
                system("/usr/local/etc/rc.d/mysql-server.sh start")
                command = "mail -s'Mysqld reset' root@example.com\n"
        }
        else
        {
                command = "mail -s'Mysqld ok' root@example.com\n"
        }

        printf("%s\ncount: %d\n%s",strftime("%c"),count,report) | command
}

then chmod +x /root/check.awk - also, make sure you have put your own email
address in place of the root@example.com that occurs twice in that script.
Also, you may just put a null command in place of the 'Mysqld ok' line if
you don't need to be alerted of positive (no action) results.

Thank you very much!