Add scripts/system_health_report.sh

scripts/system_health_report.sh

@@ -0,0 +1,110 @@
+#!/usr/bin/env bash
+#
+# Generate a concise system health report for Linux servers.
+#
+# This script is useful for incident triage, routine maintenance, and
+# collecting baseline information before deployments or infrastructure changes.
+#
+# Usage:
+#   bash system_health_report.sh
+#   bash system_health_report.sh --output /tmp/health-report.txt
+#   bash system_health_report.sh --service nginx --service docker
+#
+# Options:
+#   --output FILE       Write the report to FILE instead of stdout.
+#   --service NAME     Include systemd status for a service. Can be repeated.
+#   --help             Show this help message.
+
+set -Eeuo pipefail
+
+output_file=""
+services=()
+
+usage() {
+  sed -n '2,18p' "$0" | sed 's/^# \{0,1\}//'
+}
+
+require_command() {
+  local command_name="$1"
+  if ! command -v "$command_name" >/dev/null 2>&1; then
+    printf 'ERROR: required command not found: %s\n' "$command_name" >&2
+    exit 1
+  fi
+}
+
+section() {
+  printf '\n## %s\n' "$1"
+}
+
+collect_report() {
+  section "Host"
+  printf 'Hostname: %s\n' "$(hostname -f 2>/dev/null || hostname)"
+  printf 'Kernel: %s\n' "$(uname -srmo)"
+  printf 'Uptime: %s\n' "$(uptime -p 2>/dev/null || uptime)"
+  printf 'Date: %s\n' "$(date -Is)"
+
+  section "CPU"
+  if command -v lscpu >/dev/null 2>&1; then
+    lscpu | awk -F: '/Model name|CPU\(s\)|Thread|Core|Socket/ {gsub(/^[ \t]+/, "", $2); print $1 ": " $2}'
+  else
+    grep -m1 'model name' /proc/cpuinfo || true
+  fi
+  printf 'Load average: %s\n' "$(awk '{print $1, $2, $3}' /proc/loadavg)"
+
+  section "Memory"
+  free -h
+
+  section "Disk"
+  df -hT -x tmpfs -x devtmpfs
+
+  section "Network"
+  ip -brief address 2>/dev/null || hostname -I || true
+
+  section "Top Processes By Memory"
+  ps -eo pid,ppid,comm,%mem,%cpu --sort=-%mem | head -n 11
+
+  if [[ "${#services[@]}" -gt 0 ]]; then
+    section "Service Status"
+    for service in "${services[@]}"; do
+      printf '\nService: %s\n' "$service"
+      systemctl is-active "$service" 2>/dev/null || true
+      systemctl --no-pager --lines=5 status "$service" 2>/dev/null || true
+    done
+  fi
+}
+
+while [[ "$#" -gt 0 ]]; do
+  case "$1" in
+    --output)
+      output_file="${2:-}"
+      [[ -n "$output_file" ]] || { printf 'ERROR: --output requires a file path.\n' >&2; exit 1; }
+      shift 2
+      ;;
+    --service)
+      [[ -n "${2:-}" ]] || { printf 'ERROR: --service requires a service name.\n' >&2; exit 1; }
+      services+=("$2")
+      shift 2
+      ;;
+    --help|-h)
+      usage
+      exit 0
+      ;;
+    *)
+      printf 'ERROR: unknown option: %s\n' "$1" >&2
+      usage >&2
+      exit 1
+      ;;
+  esac
+done
+
+require_command free
+require_command df
+require_command ps
+
+if [[ -n "$output_file" ]]; then
+  collect_report > "$output_file"
+  printf 'Health report written to %s\n' "$output_file"
+else
+  collect_report
+fi
+