nagios: adding check_swaping.sh

nagios/check_swaping.sh

@@ -0,0 +1,169 @@
+#!/bin/sh
+
+
+# Little custom script to :
+# - check if the system has a heavy swap activity
+# - record swapin activity as perfdata
+#
+# For pnp4nagios' sake, create a file usually named 'check_nrpe_swaping.cfg'
+# (depends on your configuration) with the following content :
+#  DATATYPE = COUNTER,COUNTER
+
+
+# Default values
+THRESHOLD_WARNING="1000"
+THRESHOLD_CRITICAL="2000"
+VMSTAT_PREVIOUS_DATA_FILE="/tmp/.monitoring_vmstat.txt"
+
+
+# Output
+OUTPUT_EXIT_STATUS=0
+OUTPUT_DETAIL_WARNING=""
+OUTPUT_DETAIL_CRITICAL=""
+OUTPUT_PERFDATA=""
+
+PROGPATH=$( echo $0 | sed -e 's,[\\/][^\\/][^\\/]*$,,' )
+REVISION="0.1"
+
+# Stop at the first non-catched error
+set -e
+
+# Include check_range()
+. $PROGPATH/utils.sh
+# If you don't have the previous file, just comment the line and uncomment the following
+# (it's only possible because we don't use the range checking function in this script)
+#STATE_OK=0
+#STATE_WARNING=1
+#STATE_CRITICAL=2
+#STATE_UNKNOWN=3
+#STATE_DEPENDENT=4
+
+
+#
+# Help function
+#
+usage() {
+	cat <<EOF
+Usage :
+  $0 [-w warning_threshold] [-c critical_threshold] [-f vmstat_previous_data_file ]
+
+Note 1 : the script will measure the number of seconds passed since its last call and will
+         divide the measures accordingly, so write the thresholds using pages/s in mind.
+         The script will measure against pswpin and pswpout added together.
+
+Note 2 : the thresholds use the kernel page size as unit. Use 'getconf PAGESIZE' to get it if needed.
+
+Default values:
+  warning_threshold         : $THRESHOLD_WARNING
+  critical_threshold        : $THRESHOLD_CRITICAL
+  vmstat_previous_data_file : $VMSTAT_PREVIOUS_DATA_FILE
+EOF
+}
+
+#
+# Copy fresh vmstat data into the file passed as an argument
+#
+update_vmstat_previous_data_file() {
+	# Harden default files permissions to avoid some data leaks
+	umask "0077" || true
+
+	echo "# This file was written by $0 for monitoring swap activity." >"$1"
+	echo "# It can be deleted if the need arise, it will be easily recreated without too much lost." >>"$1"
+	echo "" >>"$1"
+
+	cat /proc/vmstat >> "$1"
+}
+
+#
+# Check if arg is an integer
+# (copied from jilles @ http://stackoverflow.com/questions/806906/how-do-i-test-if-a-variable-is-a-number-in-bash )
+#
+is_int() {
+	case "$1" in
+		''|*[!0-9]*) return 1;;
+		*) return 0;;
+	esac
+}
+
+
+#
+# Parameters management
+#
+while getopts hw:c:f: f; do
+	case "$f" in
+		'h')
+			usage
+			exit
+			;;
+
+		'w')
+			THRESHOLD_WARNING="$OPTARG"
+			;;
+
+		'c')
+			THRESHOLD_CRITICAL="$OPTARG"
+			;;
+
+		'f')
+			VMSTAT_PREVIOUS_DATA_FILE="$OPTARG"
+			;;
+
+		\?)
+			usage
+			exit 1
+			;;
+	esac
+done
+shift $( expr $OPTIND - 1 )
+
+# Little checks
+if ! is_int "$THRESHOLD_WARNING" || ! is_int "$THRESHOLD_CRITICAL"; then
+	echo "UNKNOWN invalid parameter : one of the threshold is not an integer."
+	exit $STATE_UNKNOWN
+fi
+
+# Check if the previous data file exists
+if [ ! -w "$VMSTAT_PREVIOUS_DATA_FILE" ]; then
+	update_vmstat_previous_data_file "$VMSTAT_PREVIOUS_DATA_FILE"
+
+	# We wait a little bit to gather some data even on the first run
+	# (or we could return an UNKNOWN ?)
+	sleep 2
+fi
+
+# Read previous data
+PREVIOUS_PSWPIN="$( grep '^pswpin' "$VMSTAT_PREVIOUS_DATA_FILE" | cut -d " " -f 2 )"
+PREVIOUS_PSWPOUT="$( grep '^pswpout' "$VMSTAT_PREVIOUS_DATA_FILE" | cut -d " " -f 2 )"
+PREVIOUS_PSWPTOTAL=$(( $PREVIOUS_PSWPIN + $PREVIOUS_PSWPOUT ))
+
+# Get time elapsed since last call
+PREVIOUS_DATA_AGE=$(( $( date +%s ) - $( stat --printf="%Y" "$VMSTAT_PREVIOUS_DATA_FILE" ) ))
+if [ "$PREVIOUS_DATA_AGE" -le "0" ]; then
+	echo "UNKNOWN: $PREVIOUS_DATA_AGE second(s) elapsed since last call."
+	exit $STATE_UNKNOWN
+fi
+
+# Update and read current data
+update_vmstat_previous_data_file "$VMSTAT_PREVIOUS_DATA_FILE"
+CURRENT_PSWPIN="$( grep '^pswpin' "$VMSTAT_PREVIOUS_DATA_FILE" | cut -d " " -f 2 )"
+CURRENT_PSWPOUT="$( grep '^pswpout' "$VMSTAT_PREVIOUS_DATA_FILE" | cut -d " " -f 2 )"
+CURRENT_PSWPTOTAL=$(( $CURRENT_PSWPIN + $CURRENT_PSWPOUT ))
+
+# Calculate the swaping rate
+PSWP_RATE=$(( ( $CURRENT_PSWPTOTAL - $PREVIOUS_PSWPTOTAL ) / $PREVIOUS_DATA_AGE ))
+
+# Generate perfdata
+OUTPUT_PERFDATA="$( printf " pswpin=%d pswpout=%d" "$CURRENT_PSWPIN" "$CURRENT_PSWPOUT" )"
+
+# Comparison
+# note: remember that numbers can be reseted to zero from time to time (64bits counter ?)
+if [ $PSWP_RATE -gt $THRESHOLD_CRITICAL ]; then
+	echo "CRITICAL swaping rate at $PSWP_RATE (limit at $THRESHOLD_CRITICAL) |$OUTPUT_PERFDATA"
+	exit $STATE_CRITICAL
+elif [ $PSWP_RATE -gt $THRESHOLD_WARNING ]; then
+	echo "WARNING swaping rate at $PSWP_RATE (limit at $THRESHOLD_WARNING) |$OUTPUT_PERFDATA"
+	exit $STATE_WARNING
+fi
+
+echo "OK swaping rate at $PSWP_RATE |$OUTPUT_PERFDATA"
+exit $STATE_OK

nagios/etc/30_nrpe-basic.cfg

@@ -2,6 +2,7 @@
 command[check_disks]=/usr/lib/nagios/plugins/check_disk -f -w 10% -c 5% -W 50% -K 5% -l -X tmpfs -X devpts -X usbfs -X nsfs -X overlay
 command[check_load]=/usr/lib/nagios/plugins/check_load -w 1,1,1 -c 3,2,2
 command[check_network_volume]=/usr/local/share/scripts-admin/nagios/check_network_volume.sh
+command[check_swaping]=/usr/local/share/scripts-admin/nagios/check_swaping.sh
 
 # Petite commande temporaire pour étudier souci neighbour table overflow
 command[check_network-neighbour-table]=/usr/local/share/scripts-admin/nagios/check_network-neighbour-table.sh