#!/bin/bash
# Monitor Ceph cluster health
# DISABLED: Ceph is not being used for storage (using NFS/CIFS/LVM instead)
# Ceph services are running but unconfigured (0 OSDs), causing false warnings

set -euo pipefail

# Check if Ceph is installed
if ! command -v ceph &>/dev/null; then
    logger -t ceph-monitor "Ceph not installed, skipping"
    exit 0
fi

# Check if Ceph storage is actually being used
CEPH_STORAGE=$(pvesm status 2>/dev/null | grep -i ceph || echo "")

if [ -z "$CEPH_STORAGE" ]; then
    # Ceph is not in use, skip monitoring
    logger -t ceph-monitor "Ceph not in use, monitoring disabled"
    exit 0
fi

# Only monitor if Ceph storage is actually configured
SEND_NTFY="/usr/local/bin/send-ntfy.sh"

# Get Ceph status
CEPH_STATUS=$(timeout 10 ceph -s 2>/dev/null || echo "FAILED")

if [ "$CEPH_STATUS" = "FAILED" ]; then
    $SEND_NTFY critical "Ceph Check Failed" "🔴 CRITICAL: Unable to get Ceph cluster status!" "skull,error"
    exit 1
fi

# Check overall health
HEALTH=$(echo "$CEPH_STATUS" | grep -oP 'health: \K\w+' || echo "UNKNOWN")

if [ "$HEALTH" = "HEALTH_ERR" ]; then
    $SEND_NTFY critical "Ceph Health Error" "🔴 CRITICAL: Ceph cluster is in HEALTH_ERR state!\n$(ceph health detail 2>/dev/null | head -3)" "skull,error,cd"
elif [ "$HEALTH" = "HEALTH_WARN" ]; then
    $SEND_NTFY warning "Ceph Health Warning" "🟡 WARNING: Ceph cluster is in HEALTH_WARN state\n$(ceph health detail 2>/dev/null | head -3)" "warning,cd"
fi

# Check for degraded PGs
DEGRADED=$(echo "$CEPH_STATUS" | grep -i degraded || echo "")
if [ -n "$DEGRADED" ]; then
    $SEND_NTFY warning "Ceph PGs Degraded" "🟡 WARNING: Ceph has degraded placement groups\n$DEGRADED" "warning,cd"
fi

logger -t ceph-monitor "Ceph health: $HEALTH"