#!/usr/bin/env bash
#
# Health watchdog for the BBE site. Probes the frontend's request-time routes and
# the backend health endpoint on their local ports. On sustained failure it
# restarts bbe-frontend and logs loudly to the journal, turning a silent multi-day
# outage (like the /events 500) into a self-healing few-minute blip.
#
# Meant to be run by bbe-site-healthcheck.timer every few minutes. State is kept
# in a tmp file so a single transient blip doesn't trigger a restart — we only act
# after FAIL_THRESHOLD consecutive failing runs.

set -uo pipefail

FRONTEND_BASE="${FRONTEND_BASE:-http://127.0.0.1:4056}"
API_BASE="${API_BASE:-http://127.0.0.1:4055}"
FAIL_THRESHOLD="${FAIL_THRESHOLD:-2}"
STATE_FILE="${STATE_FILE:-/tmp/bbe-healthcheck.fails}"
TIMEOUT="${TIMEOUT:-10}"

FRONTEND_PATHS=(/ /events /blog)

log() { echo "bbe-healthcheck: $*"; }

check() {
  local url="$1" code
  code="$(curl -sS -o /dev/null -w '%{http_code}' --max-time "${TIMEOUT}" "${url}" 2>/dev/null || echo 000)"
  if [ "${code}" = "200" ]; then
    return 0
  fi
  log "FAIL ${url} -> ${code}"
  return 1
}

failed=0
for p in "${FRONTEND_PATHS[@]}"; do
  check "${FRONTEND_BASE}${p}" || failed=1
done
check "${API_BASE}/api/health" || failed=1

if [ "${failed}" -eq 0 ]; then
  # Healthy: clear the failure counter and exit quietly.
  rm -f "${STATE_FILE}"
  log "OK (all probes 200)"
  exit 0
fi

# Unhealthy: increment consecutive-failure counter.
count=0
[ -f "${STATE_FILE}" ] && count="$(cat "${STATE_FILE}" 2>/dev/null || echo 0)"
count=$((count + 1))
echo "${count}" > "${STATE_FILE}"
log "unhealthy run ${count}/${FAIL_THRESHOLD}"

if [ "${count}" -ge "${FAIL_THRESHOLD}" ]; then
  log "threshold reached — restarting bbe-frontend.service"
  if systemctl restart bbe-frontend 2>/dev/null; then
    log "restart issued"
  else
    log "ERROR: restart failed (needs privileges? see PolicyKit/sudoers note in ops/README.md)"
  fi
  rm -f "${STATE_FILE}"
fi

exit 1