Zombie Process Check Shell Script

A rising zombie process count usually means a parent is not calling wait(). This script lists STAT Z rows, counts them, and exits non-zero when monitoring tools should page someone.

What this script does

Parses ps -eo stat,pid,ppid,cmd for state Z
Configurable alert threshold (default 5)
Prints parent PID for follow-up
Logs summary to /var/log/zombie-check.log
Nagios/monitoring friendly exit codes

Prerequisites

procps ps
Read permission on /proc
Fix requires restarting or patching parent process

Step 1: Save the script

sudo nano /usr/local/bin/zombie-check.sh
sudo chmod +x /usr/local/bin/zombie-check.sh

Step 2: Full script (scroll to read)

zombie-check.sh

#!/usr/bin/env bash
set -euo pipefail

THRESHOLD="${THRESHOLD:-5}"
LOG="/var/log/zombie-check.log"

log(){ echo "[$(date '+%F %T')] $*" | tee -a "$LOG"; }

mapfile -t zombies < <(ps -eo stat=,pid=,ppid=,cmd= | awk '$1 ~ /Z/ {print}')
count=${#zombies[@]}

log "zombie count: $count (threshold $THRESHOLD)"
if (( count > 0 )); then
  printf '%s\n' "${zombies[@]}" | head -20 | tee -a "$LOG"
fi

if (( count >= THRESHOLD )); then
  log "ALERT zombies >= threshold"
  exit 1
fi
exit 0

Scroll inside the box to read the full script.

Step 3: Configure settings

Killing zombies alone does not work — restart the parent or fix the app
Related: pgrep monitor
Load issues: load alert

Shell script checking for zombie processes

Step 4: Test manually

./zombie-check.sh

THRESHOLD=0 ./zombie-check.sh  # may alert on any zombie

Schedule with cron

sudo crontab -e

Add:

*/15 * * * * /usr/local/bin/zombie-check.sh >> /var/log/zombie-check.log 2>&1

Check Zombie Processes Shell Script Linux (ps, Alert, Cron)