Files
Linux_post_install/bin/pos-system-health
T
Your Name 0b7fdb4954 feat: pos config — interactive config editor driven by the POS_CONFIG registry
Add bin/pos-config + lib/config-ui.sh: reads the '# POS_CONFIG:' headers
across tools into a unified runtime config (~/.config/linux_post_install/
<scope>.env, chmod 600). Secret masking with show/hide, digits:/num:/url:
validation, '-' clears, blank keeps; '*plugins' expands entertainment
plugin vars from entertainment-lib.sh. pos config with no scope opens a
picker (scope list when not a TTY); completion offers scopes from the
registry at runtime.

gen-docs.sh now handles category-less tools (pos-config) instead of
emitting pos-config-config; fix set -e/pipefail abort that truncated the
header registry mid-read. install.sh ships config-ui.sh.
2026-08-09 03:57:17 -04:00

276 lines
9.8 KiB
Bash
Executable File

#!/usr/bin/env bash
set -euo pipefail
# POS: system health — Host health dashboard (disk, RAM, services, backup age, fail2ban, docker); exit 1 if any FAIL
# POS_FLAGS: --send --markdown
# POS_CONFIG: system | system.env | BACKUP_SERVICE_ROOTS=:Roots scanned by backup --service and the health backup-age check (default: /srv $HOME/srv) | HEALTH_BACKUP_MAX_AGE_DAYS=num:Max backup age in days before health warns (default 2)
# POS_CONFIG: notify | notify.env | NOTIFY_PLATFORM=:Comma-separated notify platforms (default telegram)
source "$(dirname "$0")/../lib/common.sh" 2>/dev/null || source "$(dirname "$0")/common.sh"
source "$(dirname "$0")/../lib/notify.sh" 2>/dev/null || source "$(dirname "$0")/notify.sh"
load_system_env
# Effective dynamic values (environment > system.env > defaults) shown in --help
EFF_PLATFORMS="$(notify_platforms)"
EFF_MAX_AGE="${HEALTH_BACKUP_MAX_AGE_DAYS:-2}"
EFF_ROOTS="${BACKUP_SERVICE_ROOTS:-/srv $HOME/srv}"
usage() {
cat <<EOF
Usage: pos system health [--send] [--markdown]
Host health dashboard: disk, RAM/swap, failed systemd units, backup age,
fail2ban, docker containers. Exits 1 if any check FAILs.
Flags:
--send Send the summary via the configured notify platform(s) and do
NOT print the dashboard (notification-only — for cron/timers or
wrappers like the Telegram listener that reply with output)
(uses lib/notify.sh; see NOTIFY_PLATFORM below)
--markdown Same as --send, with markdown parse mode
-h, --help Show this help
Environment (effective values):
NOTIFY_PLATFORM ${EFF_PLATFORMS}
HEALTH_BACKUP_MAX_AGE_DAYS ${EFF_MAX_AGE}
BACKUP_SERVICE_ROOTS ${EFF_ROOTS}
Loaded from ~/.config/linux_post_install/system.env (health/backup settings)
and ~/.config/linux_post_install/notify.env (NOTIFY_PLATFORM) unless already
exported. Defaults apply when neither is set.
Examples:
pos system health
pos system health --send
EOF
exit 0
}
SEND=0
MARKDOWN=0
for arg in "$@"; do
case "$arg" in
-h|--help) usage ;;
--send) SEND=1 ;;
--markdown) MARKDOWN=1 ;;
*) err "Unknown option '$arg' (see --help)" ;;
esac
done
if [ "$MARKDOWN" -eq 1 ]; then
SEND=1
fi
# --send is notification-only: suppress the dashboard output (so a wrapper
# such as the Telegram listener doesn't echo it back as a duplicate reply).
if [ "$SEND" -eq 1 ]; then
exec 3>&1
exec 1>/dev/null
fi
FAILURES=0
WARNINGS=0
REPORT=()
report() {
local st="$1" label="$2" detail="$3" line
case "$st" in
ok)
echo " ${GREEN}${BOLD}[ OK ]${RESET} ${label}: ${detail}"
printf -v line "[ OK ] %s: %s" "$label" "$detail"
;;
warn)
echo " ${YELLOW}${BOLD}[WARN]${RESET} ${label}: ${detail}"
printf -v line "[WARN] %s: %s" "$label" "$detail"
WARNINGS=$((WARNINGS + 1))
;;
fail)
echo " ${RED}${BOLD}[FAIL]${RESET} ${label}: ${detail}"
printf -v line "[FAIL] %s: %s" "$label" "$detail"
FAILURES=$((FAILURES + 1))
;;
esac
REPORT+=("$line")
}
host="$(hostname -s 2>/dev/null || echo "localhost")"
ip_public="$(curl -fsS -m 5 https://api.ipify.org 2>/dev/null || echo "unreachable")"
loadavg="$(cut -d' ' -f1-3 /proc/loadavg 2>/dev/null || echo "?")"
uptime_s="$(uptime -p 2>/dev/null || echo "?")"
section "System Health — ${host}"
echo " ${BLUE}${BOLD}date${RESET} $(date '+%Y-%m-%d %H:%M:%S')"
echo " ${BLUE}${BOLD}uptime${RESET} ${uptime_s}"
echo " ${BLUE}${BOLD}load${RESET} ${loadavg}"
echo " ${BLUE}${BOLD}ip${RESET} ${ip_public}"
echo
# ── Disk ──────────────────────────────────────────────────────
DISK_PCT_MAX=0
DISK_ISSUES=()
while read -r fs use mount; do
pct=${use%\%}
if [ "$pct" -gt "$DISK_PCT_MAX" ]; then
DISK_PCT_MAX="$pct"
fi
if [ "$pct" -ge 90 ]; then
DISK_ISSUES+=("${mount} ${pct}%")
elif [ "$pct" -ge 85 ]; then
DISK_ISSUES+=("${mount} ${pct}%")
fi
done < <(df -P -x tmpfs -x devtmpfs -x squashfs -x overlay -x efivarfs -x proc -x sysfs -x cgroup2 -x zfs -x aufs 2>/dev/null | awk 'NR>1 && $1 !~ /loop|sr[0-9]/ {print $1, $5, $6}')
if [ ${#DISK_ISSUES[@]} -gt 0 ]; then
report fail "disk" "${DISK_ISSUES[*]}"
else
report ok "disk" "max usage ${DISK_PCT_MAX}%"
fi
# ── RAM / swap ────────────────────────────────────────────────
mem=($(free -m | awk '/^Mem:/{print $2, $7}'))
swap=($(free -m | awk '/^Swap:/{print $2, $3}'))
MEM_TOTAL=${mem[0]:-0}
MEM_AVAIL=${mem[1]:-0}
MEM_PCT=100
if [ "$MEM_TOTAL" -gt 0 ]; then
MEM_PCT=$(( (MEM_TOTAL - MEM_AVAIL) * 100 / MEM_TOTAL ))
fi
SWAP_PCT=0
if [ "${swap[0]:-0}" -gt 0 ]; then
SWAP_PCT=$(( swap[1] * 100 / swap[0] ))
fi
if [ "$MEM_PCT" -ge 95 ] || [ "$SWAP_PCT" -ge 90 ]; then
report fail "ram" "mem ${MEM_PCT}% used, swap ${SWAP_PCT}%"
elif [ "$MEM_PCT" -ge 85 ]; then
report warn "ram" "mem ${MEM_PCT}% used, swap ${SWAP_PCT}%"
else
report ok "ram" "mem ${MEM_PCT}% used, swap ${SWAP_PCT}%"
fi
# ── Failed systemd units ──────────────────────────────────────
if command -v systemctl >/dev/null 2>&1; then
failed_units=()
while IFS= read -r u; do
[ -n "$u" ] && failed_units+=("${u%% *}")
done < <(systemctl --failed --plain --no-legend --no-pager 2>/dev/null || true)
if [ "${#failed_units[@]}" -gt 0 ]; then
report fail "services" "${#failed_units[@]} failed unit(s): ${failed_units[*]}"
else
report ok "services" "no failed units"
fi
else
report warn "services" "systemd not available"
fi
# ── Backup age ────────────────────────────────────────────────
MAX_AGE="${HEALTH_BACKUP_MAX_AGE_DAYS:-2}"
roots=()
if [ -n "${BACKUP_SERVICE_ROOTS:-}" ]; then
read -r -a roots <<< "$BACKUP_SERVICE_ROOTS"
else
roots=(/srv "$HOME/srv")
fi
newest=""
newest_ts=0
for root in "${roots[@]}"; do
[ -d "$root" ] || continue
while IFS= read -r line; do
ts="${line%% *}"
if [ "${ts%%.*}" -gt "$newest_ts" ]; then
newest_ts="${ts%%.*}"
newest="${line#* }"
fi
done < <(find "$root" -maxdepth 2 -type f -name '*.tar.gz.gpg' -printf '%T@ %p\n' 2>/dev/null || true)
done
if [ -z "$newest" ]; then
report warn "backup" "no backups found under ${roots[*]}"
else
now="$(date +%s)"
days=$(( (now - newest_ts) / 86400 ))
if [ "$days" -gt "$MAX_AGE" ]; then
report warn "backup" "${days}d old: ${newest}"
else
report ok "backup" "${days}d old: ${newest}"
fi
fi
# ── fail2ban ──────────────────────────────────────────────────
if command -v fail2ban-client >/dev/null 2>&1; then
if fail2ban-client ping >/dev/null 2>&1; then
banned=0
jails=$(fail2ban-client status 2>/dev/null | awk -F': *' '/Jail list/{print $2}')
for j in $jails; do
n=$(fail2ban-client status "$j" 2>/dev/null | awk -F': *' '/Total banned/{print $2}')
n="${n:-0}"
banned=$(( banned + (n > 0 ? n : 0) ))
done
report ok "fail2ban" "running, ${banned} total banned across ${jails:-0} jail(s)"
else
report warn "fail2ban" "installed but not running"
fi
fi
# ── Docker ────────────────────────────────────────────────────
if command -v docker >/dev/null 2>&1; then
bad_containers=()
running=0
while IFS= read -r name state; do
case "$state" in
restarting|unhealthy) bad_containers+=("${name}(${state})") ;;
esac
done < <(docker ps -a --format '{{.Names}} {{.State}}' 2>/dev/null || true)
if [ "${#bad_containers[@]}" -gt 0 ]; then
report fail "docker" "${#bad_containers[@]} problem container(s): ${bad_containers[*]}"
else
report ok "docker" "no unhealthy/restarting containers"
fi
fi
echo
if [ "$FAILURES" -gt 0 ]; then
verdict="${FAILURES} FAIL, ${WARNINGS} WARN"
echo " ${RED}${BOLD}VERDICT${RESET} ${verdict}"
elif [ "$WARNINGS" -gt 0 ]; then
verdict="${WARNINGS} WARN — mostly healthy"
echo " ${YELLOW}${BOLD}VERDICT${RESET} ${verdict}"
else
verdict="all good"
echo " ${GREEN}${BOLD}VERDICT${RESET} ${verdict}"
fi
if [ "$SEND" -eq 1 ]; then
exec 1>&3 3>&-
msg="POS Health — ${host}
Verdict: ${verdict}"
for line in "${REPORT[@]}"; do
msg+="
${line}"
done
if [ "$MARKDOWN" -eq 1 ]; then
notify_send "$msg" --markdown
else
notify_send "$msg"
fi
# notify_send is silent-fail by design — confirm delivery only when a
# platform sender actually exists, otherwise warn on stderr (still visible
# here since only stdout was redirected while --send ran).
delivered=0
IFS=',' read -r -a platforms <<< "$(notify_platforms)"
for p in "${platforms[@]}"; do
p="${p// /}"
[ -n "$p" ] || continue
sname="$(notify_sender_name "$p")"
if command -v "pos-communication-$sname" &>/dev/null \
|| [ -x "$(dirname "$0")/pos-communication-$sname" ] \
|| [ -x "$(dirname "$0")/../bin/pos-communication-$sname" ]; then
delivered=1
fi
done
if [ "$delivered" -eq 1 ]; then
echo "sent: POS Health — ${host} (${verdict})"
else
warn "notify: no platform sender found (NOTIFY_PLATFORM=$(notify_platforms)) — digest not delivered"
fi
fi
[ "$FAILURES" -eq 0 ]