Server health check script: disk, memory and failed units

Run this on any Debian, Ubuntu, RHEL, Rocky or AlmaLinux server for a quick “is anything wrong?” answer. It only reads information, needs no root and uses only tools that are always installed (GNU coreutils, awk, systemctl, journalctl). The exit code is 0 when everything is fine and 1 when there are warnings, so it also works from a monitoring system or a systemd timer.

The script

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

DISK_WARN=${DISK_WARN:-85}
INODE_WARN=${INODE_WARN:-85}
MEM_WARN=${MEM_WARN:-10}

problems=0
ok()   { printf '  OK    %s\n' "$*"; }
warn() { printf '  WARN  %s\n' "$*"; problems=$((problems + 1)); }

echo "Health check for $(hostname) at $(date '+%Y-%m-%d %H:%M:%S')"

echo "Disk space (warn at ${DISK_WARN}%):"
while read -r fs size used avail pct mount; do
  [[ $pct == "-" ]] && continue
  p=${pct%\%}
  if (( p >= DISK_WARN )); then
    warn "$mount is ${pct} full (${avail} free of ${size}) on $fs"
  else
    ok "$mount ${pct} used"
  fi
done < <(df -hP -x tmpfs -x devtmpfs -x squashfs -x overlay | tail -n +2)

echo "Inodes (warn at ${INODE_WARN}%):"
inode_problems=0
while read -r fs inodes iused ifree ipct mount; do
  [[ $ipct == "-" ]] && continue
  p=${ipct%\%}
  if (( p >= INODE_WARN )); then
    warn "$mount uses ${ipct} of its inodes"
    inode_problems=$((inode_problems + 1))
  fi
done < <(df -iP -x tmpfs -x devtmpfs -x squashfs -x overlay | tail -n +2)
(( inode_problems == 0 )) && ok "inode usage below ${INODE_WARN}% everywhere"

echo "Memory (warn below ${MEM_WARN}% available):"
mem_total=$(awk '/^MemTotal:/ {print $2}' /proc/meminfo)
mem_avail=$(awk '/^MemAvailable:/ {print $2}' /proc/meminfo)
mem_pct=$(( mem_avail * 100 / mem_total ))
if (( mem_pct < MEM_WARN )); then
  warn "only ${mem_pct}% of memory available ($((mem_avail / 1024)) MiB)"
else
  ok "${mem_pct}% of memory available ($((mem_avail / 1024)) MiB)"
fi
swap_total=$(awk '/^SwapTotal:/ {print $2}' /proc/meminfo)
swap_free=$(awk '/^SwapFree:/ {print $2}' /proc/meminfo)
if (( swap_total > 0 )); then
  swap_used_pct=$(( (swap_total - swap_free) * 100 / swap_total ))
  if (( swap_used_pct >= 50 )); then
    warn "swap is ${swap_used_pct}% used"
  else
    ok "swap ${swap_used_pct}% used"
  fi
fi

echo "Load:"
cpus=$(nproc)
read -r load1 load5 load15 _ < /proc/loadavg
if awk -v l="$load5" -v c="$cpus" 'BEGIN { exit !(l > c) }'; then
  warn "5-minute load $load5 is higher than CPU count $cpus"
else
  ok "load $load1 $load5 $load15 on $cpus CPUs"
fi

echo "systemd units:"
if command -v systemctl >/dev/null 2>&1; then
  failed=$(systemctl list-units --failed --no-legend --plain 2>/dev/null | awk '{print $1}')
  if [[ -n $failed ]]; then
    while read -r unit; do warn "failed unit: $unit"; done <<< "$failed"
  else
    ok "no failed units"
  fi
  if command -v journalctl >/dev/null 2>&1; then
    ok "$(journalctl --disk-usage 2>/dev/null | tail -n 1)"
  fi
else
  ok "systemctl not found, skipped"
fi

echo "Pending reboot:"
if [[ -f /var/run/reboot-required ]]; then
  warn "reboot required (/var/run/reboot-required exists)"
elif command -v needs-restarting >/dev/null 2>&1; then
  if needs-restarting -r >/dev/null 2>&1; then
    ok "no reboot needed"
  else
    warn "needs-restarting -r reports a reboot is needed"
  fi
else
  ok "no reboot marker found"
fi

echo
if (( problems > 0 )); then
  echo "RESULT: $problems warning(s)"
  exit 1
fi
echo "RESULT: all checks passed"

How to use it

sudo install -m 755 healthcheck.sh /usr/local/bin/healthcheck
healthcheck
DISK_WARN=90 MEM_WARN=5 healthcheck      # custom thresholds
healthcheck || echo "something needs attention"

What it checks

• Disk space on real filesystems (tmpfs, squashfs and overlay mounts are skipped) against DISK_WARN, default 85%.
• Inodes: a filesystem can report “No space left on device” with free space left if it ran out of inodes, usually because of millions of tiny files.
• Memory using MemAvailable from /proc/meminfo, which estimates how much memory new programs can get without swapping. “Free” memory alone is misleading because Linux uses idle RAM for cache.
• Swap usage above 50%.
• Load: the 5-minute load average compared with the CPU count.
• Failed units from systemctl list-units --failed, and how much space the journal uses.
• Pending reboot: /var/run/reboot-required on Debian and Ubuntu, or needs-restarting -r (from dnf-utils / dnf-plugins-core) on RHEL family systems, which exits 1 when a reboot is needed.

Running it on a schedule

Pair it with a systemd timer (see the timer guide) or call it from your monitoring agent. Because the output is plain text, the journal keeps a history of every run:

[Unit]
Description=Server health check

[Service]
Type=oneshot
ExecStart=/usr/local/bin/healthcheck

Official documentation

• df(1)
• systemctl(1)
• needs-restarting (dnf-plugins-core)