The CHECK_PROCS plugin
Checks system processes and does check against metrics. Default metrics is number of processes.
Supported parameters are :
- --metric/-m [METRIC]: Metric type. Valid values are: PROCS - number of processes; VSZ - virtual memory size (unix only); RSS - resident set memory size (unix only); MEM - memory usage in KB (Windows only); CPU - CPU percentage; ELAPSED - elapsed time in seconds (unix only)
- --warning/-w [THRESHOLD]: warning threshold pair.
- --critical/-c [THRESHOLD]: Critical value if metric is out of range.
- --argument-array/-a [STRING]: Only scan for processes with args that contain STRING (unix only). Use instead of ereg-argument-array.
- --ereg-argument-array [STRING]: Only scan for processes with args that contain the regex (unix only). Use instead of ereg-argument-array.
- --command/-C [STRING]: Only scan for exact matches of COMMAND (without path).
- --ppid/-p [PPID]: Only scan for children of the parent process ID indicated (unix only).
- --vsz/-v [VSZ]: Only scan for processes with VSZ (Virtual Memory Size)s higher than indicated (unix only).
- --rss/-r [RSS]: Only scan for processes with RSS (Resident Set Size) higher than indicated (unix only).
- --memory/-M [MEMORY]: Only scan for Windows processes with memory usage higher than indicated (Windows only).
Usage Example
Let us create a configuration to check warning and critical values for a particular metric
check_procs : CHECK_PROCS --warning $ARG1$ --critical $ARG2$ --metric $ARG3$
Invocation
The following invocation will raise alerts if VSZ of any processes over 50K or 100K
check_nrpe -n -H my.jnrpe.server -c check_procs -a 50000: 100000: VSZ