NAME
check_cpu - check cpu load levels and uptimes
DESCRIPTION
This is a client check that is run by spong-client that checks the
client host's cpu load levels and uptime. check_cpu used the output
of the uptime command by checking the the second load average number
(a.k.a 5 minute load average). The load average is checked against the
$CPUWARN
and $CPUCRIT levels. If the load level exceeds $CPUWARN,
it triggers a warning - yellow, and if the load level exceeds $CPUCRIT,
it triggers an alert - red.
The uptime is client host is also checked. If the uptime of the system
is less then one hour, a warning is triggered to show that the system
has been restarted recently.
OUTPUT RETURNED
In the status summary field, the system uptime, 5 minute load average,
current number of users logged in and the total number job running on the
system are returned.
If the system ps command is capable of it, the top 10 processes, sorted
by cpu percentage, are returned in the status detailed message field.
CONFIGURATION
Here is a list of all of the configuration variables used by check_cpu.
$CPUWARN, $CPUCRIT
A number inidicating the CPU load that triggers a problem ($CPUWARN
triggers warning - yellow, and $CPUCRIT triffers alerts - red).
$UPTIME
An OS specific variable that contains the full path to the uptime
command and any other commands needed to get the uptime output needed for
parsing.
$PS
An OS specific variable that contains the full path to the ps
command with the paramters needed to the ps command output to display in
the status detailed messages field.
FILES
-
/usr/local/etc/spong/spong.conf
-
Configuration file. This contains variables that detail spong and OS specific
definitions used by spong-client. This file also contain variables describing
threshold levels. See spong.conf for additional documentation.
AUTHOR
Stephen L Johnson (stephen.johnson@mail.state.ar.us)
or (sjohnson@monsters.org),
Unix System Administator, DIS - State of Arkansas