Differences

This shows you the differences between two versions of the page.

--- linux_wiki:xymon_cpu_load_threshold_calc [2017/01/09 15:32]
billdozor [The Script]
+++ linux_wiki:xymon_cpu_load_threshold_calc [2019/05/25 23:50]
@@ Line 1: / Line 1: @@
-====== Xymon CPU Load Threshold Calc ======
-**General Information**
-Posted to the Xymon Community at: https://wiki.xymonton.org/doku.php/monitors:cpu-load-calc
-Calculate (and set) a Xymon client's cpu load thresholds, based off of the client's reported number of processors from the hostdata files.
-This is mean to run on the Xymon Server periodically with cron.
-It allows for far more useful load monitoring than arbitrarily setting a generic load or spending a lot of time editing config files for each system.
-**Checklist**
-  * Xymon server installed/configured
-  * Xymon clients checking in
-----
-====== Installation ======
-Installation instructions.
-===== Xymon Client side =====
-No client modifications required.
-----
-===== Xymon Server side =====
-  - Install the 'bc' package (allows bash string to floating point number conversion)
-    - Enterprise Linux 6/7<code bash>yum install bc</code>
-  - Create the cpu-load-calc.sh script somewhere such as: /root/scripts/cpu-load-calc.sh
-    - See source below for contents
-    - Edit the "Customize Here" section
-      - Edit multipliers if desired(number of procs * number for warning and critical CPU load thresholds)
-  - Create the auto load directory on the Xymon server
-    - Default: /etc/xymon/analysis.d/auto-cpuload.d
-  - Add the auto load directory to the Xymon main analysis.cfg file so it is included
-    - Default: "directory /etc/xymon/analysis.d/auto-cpuload.d" to /etc/xymon/analysis.cfg
-  - Set the default load in /etc/xymon/analysis.cfg to a super low warning level of "0.1" and a critical of something very high like "100.0".
-    - That way, a hostdata file is generated for a system that has not had its load auto calculated. (and you don't get alert emails/texts as its only at warning)
-  - Setup the script to auto run via cron (-v is verbose output)
-    - Example<code bash>/etc/cron.hourly/xymon-calc-cpuload.sh
-#!/bin/bash
-# Description: Execute xymon script to auto calculate cpu load from hostdata
-/root/scripts/cpu-load-calc.sh -v &> /root/scripts/cpu-load-calc.log</code>
-----
-====== The Script ======
-<code bash cpu-load-calc.sh>
-#!/bin/bash
-# Title: cpu-load-calc.sh
-# Description: Calculate a xymon client's cpu load (Run on Xymon Server periodically with cron)
-# Dependency: Requires 'bc' package
-#=======================
-# Customize Here
-#=======================
-# Warning and Critical Load Multipliers (num of procs * multiplier)
-load_warn_multiplier=1.0
-load_crit_multiplier=1.5
-# Directory to save auto load thresholds
-auto_load_dir="/etc/xymon/analysis.d/auto-cpuload.d"
-# Xymon server's hostdata directory
-xymon_hostdata_dir="/var/lib/xymon/hostdata"
-# Xymon server's main analysis config file
-xymon_analysis_cfg="/etc/xymon/analysis.cfg"
-#=======================
-# End of Customize
-#=======================
-#=======================
-# Pre-Run Error Checking
-#=======================
-## Dependency Check ##
-which bc &> /dev/null
-if [[ $? -eq 1 ]]; then
-  echo ">> Error! Dependent package 'bc' (byte code) not detected. Exiting..."
-  exit 1
-fi
-## Does the Auto Load Directory exist?
-if [[ ! -d ${auto_load_dir} ]]; then
-  echo ">> Error! The directory (${auto_load_dir}) does not exist or is not a directory. Exiting..."
-  exit
-fi
-## Write Access Check
-touch ${auto_load_dir}/testfile &> /dev/null
-if [[ $? -eq 1 ]]; then
-  echo ">> Error! User '$(whoami)' does not have write access to ${auto_load_dir}! Exiting..."
-  exit 1
-else
-  rm -f ${auto_load_dir}/testfile &> /dev/null
-fi
-## Check if the auto_load_dir is included in main analysis config file
-grep "^directory ${auto_load_dir}" ${xymon_analysis_cfg} &> /dev/null
-if [[ $? -eq 1 ]]; then
-  echo -e ">> Warning! Auto load directory (${auto_load_dir}) is not included in ${xymon_analysis_cfg}. Continuing, but auto CPU load settings will not take affect until 'directory ${auto_load_dir}' is added to ${xymon_analysis_cfg}.\n"
-fi
-#=======================
-# End of Pre-Run Error Checking
-#=======================
-#===============================
-# Functions; Main starts after
-#===============================
-function show_usage
-{
-  echo -e "\n####==== Xymon Client Auto Load Thresholds ====####"
-  echo -e "\nDescripton: Calculate a xymon client's cpu load."
-  echo -e "\n--Usage"
-  echo -e "$0      => No arguments, configure with no verbosity."
-  echo -e "$0 -v   => Verbose output."
-  echo -e "$0 -r   => Refresh CPU load data (force hostdata update)."
-  echo -e "$0 -h   => Display usage."
-}
-# Force snapshots of hostdata
-function force_hostdata
-{
-  # Use node name passed as argument
-  node_name=${1}
-  # Lie to Xymon that the node's cpu is green, then yellow, forcing a hostdata snapshot
-  xymon 127.0.0.1 "status ${node_name}.cpu green $(date)"
-  xymon 127.0.0.1 "status ${node_name}.cpu yellow $(date)"
-}
-#=======================
-# Get Script Arguments
-#=======================
-# Reset POSIX variable in case it has been used previously in this shell
-OPTIND=1
-# By default, no verbose output
-verbose_output="no"
-refresh_cpus="no"
-while getopts "hrv" opt; do
-  case "${opt}" in
-    h) # -h (help) argument
-      show_usage
-      exit 0
-    ;;
-    r) # -r (refersh cpus) argument
-      refresh_cpus="yes"
-    ;;
-    v) # -v (verbose) argument
-      verbose_output="yes"
-    ;;
-    *) # invalid argument
-      show_usage
-      exit 0
-    ;;
-  esac
-done
-#=======================
-# Main Program
-#=======================
-echo -e "== Xymon Client Auto Load Thresholds =="
-echo -e "Load Warning Multiplier: ${load_warn_multiplier}"
-echo -e "Load Critical Multiplier: ${load_crit_multiplier}"
-echo -e "Saving configs to: ${auto_load_dir}"
-# For each node reporting host data
-for node in $(ls ${xymon_hostdata_dir}); do
-  if [[ ${verbose_output} == "yes" ]]; then
-    echo -e "\n>> Working on node: ${node}"
-  fi
-  if [[ ${refresh_cpus} == "yes" ]]; then
-    if [[ ${verbose_output} == "yes" ]]; then
-      echo -e "\n-> Refreshing hostdata..."
-    fi
-    # Force an update of hostdata
-    force_hostdata ${node}
-  fi
-  # Get the number of procs reported from node's most recent host data file
-  node_num_procs="$(cat ${xymon_hostdata_dir}/${node}/$(ls -tr ${xymon_hostdata_dir}/${node}/ | tail -1) | awk '/nproc/ { getline; print }')"
-  # If node_num_procs is empty or not a number, move to the next node
-  if [[ -z ${node_num_procs} || ! ${node_num_procs} =~ [0-9][0-9]* ]]; then
-    # Did not find 'nproc' in the host data file or no number from nproc returned
-    if [[ ${verbose_output} == "yes" ]]; then
-      echo "-> Warning! Could not find 'nproc' in ${node}'s host data file or no number returned. Skipping..."
-    fi
-    continue
-  fi
-  # Calculate the warning and critical load thresholds (normalize as a floating point with bc)
-  load_warning=$(echo "${node_num_procs} * ${load_warn_multiplier}" | bc)
-  load_critical=$(echo "${node_num_procs} * ${load_crit_multiplier}" | bc)
-  if [[ ${verbose_output} == "yes" ]]; then
-    echo -e "-> Number of Procs: ${node_num_procs}"
-    echo -e "-> Warning at: ${load_warning}"
-    echo -e "-> Critical at: ${load_critical}"
-    echo -e "-> Creating node analysis drop in file..."
-  fi
-  # Create analysis drop in file
-  echo "# ${node}'s CPU Load Thresholds (Warning Critical)" > ${auto_load_dir}/${node}.cfg
-  echo "HOST=${node}" >> ${auto_load_dir}/${node}.cfg
-  echo "  LOAD  ${load_warning}  ${load_critical}" >> ${auto_load_dir}/${node}.cfg
-done
-echo -e "\n== Auto Load Thresholds Complete =="
-exit 0
-</code>
-----