#!/usr/bin/env bash
set -u

MAIL_TO=${MAIL_TO:-root}
WARN_PCT=${WARN_PCT:-90}
CRIT_PCT=${CRIT_PCT:-95}
HOST=$(hostname -s 2>/dev/null || hostname)
SEND_MAIL=1

if [ "${1:-}" = "--no-mail" ]; then
  SEND_MAIL=0
fi

problems=()

add_problem() {
  problems+=("$1")
}

check_mounts() {
  if [ ! -r /etc/fstab ]; then
    printf 'MOUNT CONFIG: /etc/fstab is not readable\n'
    return
  fi

  awk '
    NF && $1 !~ /^#/ && $2 != "none" && $3 != "swap" {
      print $1 "\t" $2 "\t" $3 "\t" $4
    }
  ' /etc/fstab | while IFS=$'\t' read -r src target fstype options; do
    case ",${options}," in
      *,noauto,*) continue ;;
    esac

    if ! findmnt --mountpoint "$target" >/dev/null 2>&1; then
      printf 'MOUNT MISSING: %s source=%s type=%s options=%s\n' \
        "$target" "$src" "$fstype" "$options"
    fi
  done
}

check_disk_usage() {
  declare -A seen_fs=()

  findmnt -rn -o TARGET,FSTYPE,SOURCE | while read -r target fstype source; do
    case "$fstype" in
      autofs|binfmt_misc|bpf|cgroup|cgroup2|configfs|debugfs|devpts|devtmpfs|efivarfs|fusectl|hugetlbfs|mqueue|nfs|nfs4|nsfs|proc|pstore|rpc_pipefs|securityfs|squashfs|sysfs|tmpfs|tracefs)
        continue
        ;;
    esac

    df_line=$(df -P "$target" 2>/dev/null | awk 'NR == 2 {print}')
    [ -n "$df_line" ] || continue

    df_source=$(awk '{print $1}' <<<"$df_line")
    [ -n "$df_source" ] || df_source=$source
    if [ -n "${seen_fs[$df_source]:-}" ]; then
      continue
    fi
    seen_fs[$df_source]=1

    usage=$(awk '{gsub(/%/, "", $5); print $5}' <<<"$df_line")
    [ -n "$usage" ] || continue
    [[ "$usage" =~ ^[0-9]+$ ]] || continue

    if [ "$usage" -ge "$CRIT_PCT" ]; then
      printf 'DISK CRITICAL: %s usage=%s%% threshold=%s%% source=%s type=%s\n' \
        "$target" "$usage" "$CRIT_PCT" "$source" "$fstype"
    elif [ "$usage" -ge "$WARN_PCT" ]; then
      printf 'DISK WARNING: %s usage=%s%% threshold=%s%% source=%s type=%s\n' \
        "$target" "$usage" "$WARN_PCT" "$source" "$fstype"
    fi
  done
}

check_zfs() {
  command -v zpool >/dev/null 2>&1 || return

  pools=$(zpool list -H -o name 2>/dev/null || true)
  [ -n "$pools" ] || return

  status=$(zpool status -x 2>&1)
  if ! printf '%s\n' "$status" | grep -Eq '^all pools are healthy$'; then
    printf 'ZFS HEALTH: zpool status -x reported a problem\n%s\n' "$status"
  fi
}

while IFS= read -r line; do
  [ -n "$line" ] && add_problem "$line"
done < <(check_mounts)

while IFS= read -r line; do
  [ -n "$line" ] && add_problem "$line"
done < <(check_disk_usage)

zfs_output=$(check_zfs)
if [ -n "$zfs_output" ]; then
  add_problem "$zfs_output"
fi

if [ "${#problems[@]}" -eq 0 ]; then
  exit 0
fi

{
  printf 'Storage alerts on %s\n\n' "$HOST"
  printf 'Thresholds: warning=%s%% critical=%s%%\n\n' "$WARN_PCT" "$CRIT_PCT"
  printf '%s\n' "${problems[@]}"
} > /tmp/cluster-storage-alerts.$$

cat /tmp/cluster-storage-alerts.$$

if [ "$SEND_MAIL" -eq 1 ]; then
  if command -v mail >/dev/null 2>&1; then
    mail -s "Storage alert on $HOST" "$MAIL_TO" < /tmp/cluster-storage-alerts.$$
  elif command -v sendmail >/dev/null 2>&1; then
    {
      printf 'To: %s\n' "$MAIL_TO"
      printf 'Subject: Storage alert on %s\n\n' "$HOST"
      cat /tmp/cluster-storage-alerts.$$
    } | sendmail -t
  else
    rm -f /tmp/cluster-storage-alerts.$$
    echo "No mail or sendmail command found" >&2
    exit 2
  fi
fi

rm -f /tmp/cluster-storage-alerts.$$
if [ "$SEND_MAIL" -eq 0 ]; then
  exit 1
fi
exit 0