#!/usr/bin/env bash
# AWOOOI P3 controlled release gate.
# Read-only: this script never starts, stops, restarts, deletes, or modifies services.

set -uo pipefail

ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
cd "$ROOT_DIR"

SSH_BATCH_MODE=${SSH_BATCH_MODE:-yes}
SSH_OPTS=(-o BatchMode="$SSH_BATCH_MODE" -o ConnectTimeout=6)
NO_COLOR=0
SKIP_COLD_START_GATE=0
LOAD5_PER_CORE_LIMIT="${LOAD5_PER_CORE_LIMIT:-1.0}"
LOAD15_PER_CORE_LIMIT="${LOAD15_PER_CORE_LIMIT:-1.0}"
JOB_CONTAINER_CPU_LIMIT="${JOB_CONTAINER_CPU_LIMIT:-1.0}"
TEXTFILE_MAX_AGE_SECONDS="${TEXTFILE_MAX_AGE_SECONDS:-300}"

usage() {
  cat <<'USAGE'
Usage: bash scripts/reboot-recovery/p3-controlled-release-gate.sh [options]

Options:
  --skip-cold-start-gate  Do not run the full P0/P1/P2 read-only gate first.
  --no-color              Disable ANSI colors.
  -h, --help              Show this help.

Environment overrides:
  LOAD5_PER_CORE_LIMIT=1.0
  LOAD15_PER_CORE_LIMIT=1.0
  JOB_CONTAINER_CPU_LIMIT=1.0
  TEXTFILE_MAX_AGE_SECONDS=300
USAGE
}

while [ "$#" -gt 0 ]; do
  case "$1" in
    --skip-cold-start-gate)
      SKIP_COLD_START_GATE=1
      ;;
    --no-color)
      NO_COLOR=1
      ;;
    -h|--help)
      usage
      exit 0
      ;;
    *)
      echo "Unknown argument: $1" >&2
      usage >&2
      exit 64
      ;;
  esac
  shift
done

if [ "$NO_COLOR" = "1" ]; then
  RED=""
  GREEN=""
  YELLOW=""
  BLUE=""
  NC=""
else
  RED=$'\033[0;31m'
  GREEN=$'\033[0;32m'
  YELLOW=$'\033[1;33m'
  BLUE=$'\033[0;34m'
  NC=$'\033[0m'
fi

PASS=0
WARN=0
FAIL=0

section() {
  printf "\n%s=== %s ===%s\n" "$BLUE" "$1" "$NC"
}

ok() {
  PASS=$((PASS + 1))
  printf "%sOK%s %s\n" "$GREEN" "$NC" "$*"
}

warn() {
  WARN=$((WARN + 1))
  printf "%sWARN%s %s\n" "$YELLOW" "$NC" "$*"
}

blocked() {
  FAIL=$((FAIL + 1))
  printf "%sBLOCKED%s %s\n" "$RED" "$NC" "$*"
}

ssh_cmd() {
  local target="$1"
  local cmd="$2"
  ssh "${SSH_OPTS[@]}" "$target" "$cmd"
}

float_le() {
  awk -v a="$1" -v b="$2" 'BEGIN { exit !(a <= b) }'
}

check_cold_start_gate() {
  section "P0/P1/P2 cold-start gate"
  if [ "$SKIP_COLD_START_GATE" -eq 1 ]; then
    warn "cold-start gate skipped by operator option"
    return
  fi

  SSH_BATCH_MODE=yes bash scripts/reboot-recovery/full-stack-cold-start-check.sh --monitor-read-only --no-color --watch --interval 1 --max-attempts 1 >/tmp/awoooi-p3-cold-start-gate.log 2>&1
  local rc=$?
  local summary blocked_count warn_count
  summary=$(grep -E '^PASS=[0-9]+ WARN=[0-9]+ BLOCKED=[0-9]+' /tmp/awoooi-p3-cold-start-gate.log | tail -1 || true)
  blocked_count=$(awk -F'BLOCKED=' '/^PASS=/ {print $2}' <<<"$summary")
  warn_count=$(awk -F'WARN=' '/^PASS=/ {split($2,a," "); print a[1]}' <<<"$summary")

  if [ "$rc" -eq 0 ]; then
    ok "cold-start gate is GREEN"
  elif [ "${blocked_count:-1}" = "0" ]; then
    warn "cold-start gate is DEGRADED but not blocked: ${summary:-summary unavailable}"
  else
    blocked "cold-start gate has blocked items: ${summary:-summary unavailable}; see /tmp/awoooi-p3-cold-start-gate.log"
  fi
}

check_host_load() {
  local label="$1"
  local target="$2"
  local out load5 load15 cores load5_per_core load15_per_core

  section "$label load gate"
  if ! out=$(ssh_cmd "$target" 'read _ load5 load15 _ < /proc/loadavg; cores=$(nproc); awk -v l5="$load5" -v l15="$load15" -v c="$cores" "BEGIN {printf \"LOAD5 %.4f LOAD15 %.4f CORES %d LOAD5_PER_CORE %.6f LOAD15_PER_CORE %.6f\\n\", l5, l15, c, l5/c, l15/c}"' 2>&1); then
    blocked "$label load check unavailable"
    echo "$out"
    return
  fi
  echo "$out"
  load5_per_core=$(awk '/LOAD5_PER_CORE/ {for (i=1;i<=NF;i++) if ($i=="LOAD5_PER_CORE") print $(i+1)}' <<<"$out")
  load15_per_core=$(awk '/LOAD15_PER_CORE/ {for (i=1;i<=NF;i++) if ($i=="LOAD15_PER_CORE") print $(i+1)}' <<<"$out")
  if float_le "$load5_per_core" "$LOAD5_PER_CORE_LIMIT"; then
    ok "$label load5/core <= $LOAD5_PER_CORE_LIMIT"
  else
    blocked "$label load5/core too high for P3 release"
  fi
  if float_le "$load15_per_core" "$LOAD15_PER_CORE_LIMIT"; then
    ok "$label load15/core <= $LOAD15_PER_CORE_LIMIT"
  else
    blocked "$label load15/core too high for P3 release"
  fi
}

check_textfiles() {
  section "textfile freshness"
  local out
  if out=$(ssh_cmd "wooo@192.168.0.110" '
now=$(date +%s)
for f in /home/wooo/node_exporter_textfiles/docker_stats.prom /home/wooo/node_exporter_textfiles/systemd_units.prom /home/wooo/node_exporter_textfiles/storage_health.prom /home/wooo/node_exporter_textfiles/backup_health.prom /home/wooo/node_exporter_textfiles/cold_start_recovery.prom; do
  if [ -f "$f" ]; then
    echo "$(basename "$f") $((now - $(stat -c %Y "$f")))"
  else
    echo "$(basename "$f") missing"
  fi
done
' 2>&1); then
    while read -r file age; do
      [ -n "${file:-}" ] || continue
      max_age="$TEXTFILE_MAX_AGE_SECONDS"
      [ "$file" = "cold_start_recovery.prom" ] && max_age=900
      [ "$file" = "backup_health.prom" ] && max_age=900
      if [ "$age" = "missing" ]; then
        blocked "110 $file missing"
      elif [ "$age" -le "$max_age" ]; then
        ok "110 $file fresh age=${age}s"
      else
        blocked "110 $file stale age=${age}s"
      fi
    done <<<"$out"
  else
    blocked "110 textfile freshness check unavailable"
    echo "$out"
  fi

  if out=$(ssh_cmd "ollama@192.168.0.188" '
now=$(date +%s)
for f in /home/ollama/node_exporter_textfiles/docker_stats.prom /home/ollama/node_exporter_textfiles/docker_restart_count.prom /home/ollama/node_exporter_textfiles/storage_health.prom /home/ollama/node_exporter_textfiles/backup.prom /home/ollama/node_exporter_textfiles/backup_health.prom; do
  if [ -f "$f" ]; then
    echo "$(basename "$f") $((now - $(stat -c %Y "$f")))"
  else
    echo "$(basename "$f") missing"
  fi
done
' 2>&1); then
    while read -r file age; do
      [ -n "${file:-}" ] || continue
      max_age="$TEXTFILE_MAX_AGE_SECONDS"
      [ "$file" = "backup.prom" ] && max_age=90000
      [ "$file" = "backup_health.prom" ] && max_age=900
      if [ "$age" = "missing" ]; then
        blocked "188 $file missing"
      elif [ "$age" -le "$max_age" ]; then
        ok "188 $file fresh age=${age}s"
      else
        blocked "188 $file stale age=${age}s"
      fi
    done <<<"$out"
  else
    blocked "188 textfile freshness check unavailable"
    echo "$out"
  fi
}

check_backup_health() {
  section "backup health gate"
  local label target file out stale missing_cron missing_script failed_count integrity_stale
  for spec in \
    "110|wooo@192.168.0.110|/home/wooo/node_exporter_textfiles/backup_health.prom" \
    "188|ollama@192.168.0.188|/home/ollama/node_exporter_textfiles/backup_health.prom"; do
    label=${spec%%|*}
    target=${spec#*|}
    target=${target%%|*}
    file=${spec##*|}
    if ! out=$(ssh_cmd "$target" "
if [ ! -f '$file' ]; then
  echo 'BACKUP_HEALTH missing'
  exit 0
fi
awk '
  /^awoooi_backup_job_fresh/ {total += 1; stale += (\$2 == 0)}
  /^awoooi_backup_job_configured/ {missing_cron += (\$2 == 0)}
  /^awoooi_backup_script_present/ {missing_script += (\$2 == 0)}
  /^awoooi_backup_last_run_failed_count/ {failed += \$2}
  /^awoooi_backup_integrity_fresh/ {integrity_total += 1; integrity_stale += (\$2 == 0)}
  END {printf \"BACKUP_HEALTH total=%d stale=%d missing_cron=%d missing_script=%d failed_count=%d integrity_total=%d integrity_stale=%d\\n\", total, stale, missing_cron, missing_script, failed, integrity_total, integrity_stale}
' '$file'
" 2>&1); then
      blocked "$label backup health check unavailable"
      echo "$out"
      continue
    fi
    echo "$label $out"
    if grep -q "BACKUP_HEALTH missing" <<<"$out"; then
      blocked "$label backup_health.prom missing"
      continue
    fi
    stale=$(awk -F'stale=' '{split($2,a," "); print a[1]+0}' <<<"$out")
    missing_cron=$(awk -F'missing_cron=' '{split($2,a," "); print a[1]+0}' <<<"$out")
    missing_script=$(awk -F'missing_script=' '{split($2,a," "); print a[1]+0}' <<<"$out")
    failed_count=$(awk -F'failed_count=' '{split($2,a," "); print a[1]+0}' <<<"$out")
    integrity_stale=$(awk -F'integrity_stale=' '{split($2,a," "); print a[1]+0}' <<<"$out")
    [ "$stale" -eq 0 ] && ok "$label expected backups are fresh" || blocked "$label expected backup jobs are stale"
    [ "$missing_cron" -eq 0 ] && ok "$label expected backup crons are configured" || blocked "$label expected backup cron config missing"
    [ "$missing_script" -eq 0 ] && ok "$label expected backup scripts are present" || blocked "$label expected backup scripts missing"
    if [ "$label" = "110" ]; then
      [ "$integrity_stale" -eq 0 ] && ok "110 backup integrity and restore drill are fresh" || blocked "110 backup integrity or restore drill stale"
      [ "$failed_count" -eq 0 ] && ok "110 latest aggregate backup had no failed components" || warn "110 latest aggregate backup still records failed components; rerun backup-all after fixes"
    fi
  done
}

check_storage_health() {
  section "storage health gate"
  local label target file out root_readonly current_errors previous_errors fsck_errors
  for spec in \
    "110|wooo@192.168.0.110|/home/wooo/node_exporter_textfiles/storage_health.prom" \
    "188|ollama@192.168.0.188|/home/ollama/node_exporter_textfiles/storage_health.prom"; do
    label=${spec%%|*}
    target=${spec#*|}
    target=${target%%|*}
    file=${spec##*|}
    if ! out=$(ssh_cmd "$target" "
if [ ! -f '$file' ]; then
  echo 'STORAGE_HEALTH missing'
  exit 0
fi
awk '
  /^awoooi_host_root_filesystem_readonly/ {root += \$2}
  /^awoooi_host_storage_error_count/ && /boot=\"current\"/ {current += \$2}
  /^awoooi_host_storage_error_count/ && /boot=\"previous\"/ {previous += \$2}
  /^awoooi_host_storage_error_count/ && /boot=\"last-fsck-log\"/ {fsck += \$2}
  END {printf \"STORAGE_HEALTH root_readonly=%d current=%d previous=%d fsck=%d\\n\", root, current, previous, fsck}
' '$file'
" 2>&1); then
      blocked "$label storage health check unavailable"
      echo "$out"
      continue
    fi
    echo "$label $out"
    if grep -q "STORAGE_HEALTH missing" <<<"$out"; then
      blocked "$label storage_health.prom missing"
      continue
    fi
    root_readonly=$(awk -F'root_readonly=' '{split($2,a," "); print a[1]+0}' <<<"$out")
    current_errors=$(awk -F'current=' '{split($2,a," "); print a[1]+0}' <<<"$out")
    previous_errors=$(awk -F'previous=' '{split($2,a," "); print a[1]+0}' <<<"$out")
    fsck_errors=$(awk -F'fsck=' '{split($2,a," "); print a[1]+0}' <<<"$out")
    [ "$root_readonly" -eq 0 ] && ok "$label root filesystem is writable" || blocked "$label root filesystem is read-only"
    [ "$current_errors" -eq 0 ] && ok "$label current boot has no storage error evidence" || blocked "$label current boot has storage error evidence"
    [ "$previous_errors" -eq 0 ] && ok "$label previous boot has no storage error evidence" || warn "$label previous boot has storage error evidence; keep fsck/backup follow-up open"
    [ "$fsck_errors" -eq 0 ] && ok "$label fsck logs have no retained error evidence" || warn "$label fsck logs retain error evidence; verify offline fsck/backup status"
  done
}

check_runner_guardrails() {
  section "runner/CD guardrails"
  local out bad
if ! out=$(ssh_cmd "wooo@192.168.0.110" '
bad=0
for u in awoooi-direct-runner-open.service awoooi-direct-runner.service gitea-act-runner-host.service gitea-act-runner-awoooi-controlled.service gitea-awoooi-controlled-runner.service gitea-act-runner-awoooi-open.service; do
  load=$(systemctl show "$u" -p LoadState --value 2>/dev/null || true)
  unitfile=$(systemctl show "$u" -p UnitFileState --value 2>/dev/null || true)
  active=$(systemctl show "$u" -p ActiveState --value 2>/dev/null || true)
  unit_ok=0
  if [ "$load" = "masked" ] && [ "$unitfile" = "masked" ] && [ "$active" = "inactive" ]; then
    unit_ok=1
  fi
  echo "RUNNER_FAILCLOSED_UNIT $u load=$load unitfile=$unitfile active=$active ok=$unit_ok"
  [ "$unit_ok" = "1" ] || bad=1
done
cd_lane_load=$(systemctl show awoooi-cd-lane.service -p LoadState --value 2>/dev/null || true)
cd_lane_unitfile=$(systemctl show awoooi-cd-lane.service -p UnitFileState --value 2>/dev/null || true)
cd_lane_active=$(systemctl show awoooi-cd-lane.service -p ActiveState --value 2>/dev/null || true)
cd_lane_execstart=$(systemctl show awoooi-cd-lane.service -p ExecStart --value 2>/dev/null || true)
cd_lane_sentinel=missing
[ -e /run/awoooi-cd-lane-enabled ] && cd_lane_sentinel=present
cd_lane_capacity_ok=0
cd_lane_labels_ok=0
if grep -Eq "^[[:space:]]+capacity:[[:space:]]*1[[:space:]]*$" /home/wooo/awoooi-cd-lane/config.yaml 2>/dev/null; then
  cd_lane_capacity_ok=1
fi
if grep -q "awoooi-ubuntu:docker://192.168.0.110:5000/awoooi/ci-runner:act-22.04" /home/wooo/awoooi-cd-lane/config.yaml 2>/dev/null \
  && grep -q "awoooi-host:host" /home/wooo/awoooi-cd-lane/config.yaml 2>/dev/null \
  && ! grep -Eq "^[[:space:]]+- \".*(ubuntu-latest|stockplatform|headless|playwright)" /home/wooo/awoooi-cd-lane/config.yaml 2>/dev/null; then
  cd_lane_labels_ok=1
fi
cd_lane_binary_kind=$(file -b /home/wooo/awoooi-cd-lane/awoooi_cd_lane 2>/dev/null || echo missing)
cd_lane_binary_elf=0
echo "$cd_lane_binary_kind" | grep -qi "ELF" && cd_lane_binary_elf=1
cd_lane_process_count=$(pgrep -f "^/home/wooo/awoooi-cd-lane/awoooi_cd_lane" 2>/dev/null | wc -l | tr -d " ")
cd_lane_ok=0
cd_lane_mode=blocked
if [ "$cd_lane_active" = "inactive" ] \
  && [ "$cd_lane_sentinel" = "missing" ] \
  && [ "$cd_lane_binary_elf" = "0" ] \
  && [ "$cd_lane_process_count" = "0" ] \
  && { { [ "$cd_lane_load" = "masked" ] && [ "$cd_lane_unitfile" = "masked" ]; } || echo "$cd_lane_execstart" | grep -q "/bin/false"; }; then
  cd_lane_ok=1
  cd_lane_mode=failclosed
elif [ "$cd_lane_sentinel" = "present" ] && [ "$cd_lane_active" = "active" ] && [ "$cd_lane_capacity_ok" = "1" ] && [ "$cd_lane_labels_ok" = "1" ] && [ "$cd_lane_binary_elf" = "1" ]; then
  cd_lane_ok=1
  cd_lane_mode=controlled_open
fi
echo "CD_LANE_CONTROLLED mode=$cd_lane_mode load=$cd_lane_load unitfile=$cd_lane_unitfile active=$cd_lane_active sentinel=$cd_lane_sentinel capacity=$cd_lane_capacity_ok labels=$cd_lane_labels_ok binary_elf=$cd_lane_binary_elf process_count=$cd_lane_process_count ok=$cd_lane_ok"
cd_lane_drain_load=$(systemctl show awoooi-cd-lane-drain.service -p LoadState --value 2>/dev/null || true)
cd_lane_drain_unitfile=$(systemctl show awoooi-cd-lane-drain.service -p UnitFileState --value 2>/dev/null || true)
cd_lane_drain_active=$(systemctl show awoooi-cd-lane-drain.service -p ActiveState --value 2>/dev/null || true)
cd_lane_drain_cpu_accounting=$(systemctl show awoooi-cd-lane-drain.service -p CPUAccounting --value 2>/dev/null || true)
cd_lane_drain_cpu_quota=$(systemctl show awoooi-cd-lane-drain.service -p CPUQuotaPerSecUSec --value 2>/dev/null || true)
cd_lane_drain_memory_accounting=$(systemctl show awoooi-cd-lane-drain.service -p MemoryAccounting --value 2>/dev/null || true)
cd_lane_drain_memory_max=$(systemctl show awoooi-cd-lane-drain.service -p MemoryMax --value 2>/dev/null || true)
cd_lane_drain_tasks_accounting=$(systemctl show awoooi-cd-lane-drain.service -p TasksAccounting --value 2>/dev/null || true)
cd_lane_drain_tasks_max=$(systemctl show awoooi-cd-lane-drain.service -p TasksMax --value 2>/dev/null || true)
cd_lane_drain_limits_ok=0
if [ "$cd_lane_drain_cpu_accounting" = "yes" ] \
  && [ -n "$cd_lane_drain_cpu_quota" ] && [ "$cd_lane_drain_cpu_quota" != "infinity" ] \
  && [ "$cd_lane_drain_memory_accounting" = "yes" ] \
  && [ -n "$cd_lane_drain_memory_max" ] && [ "$cd_lane_drain_memory_max" != "infinity" ] \
  && [ "$cd_lane_drain_tasks_accounting" = "yes" ] \
  && [ -n "$cd_lane_drain_tasks_max" ] && [ "$cd_lane_drain_tasks_max" != "infinity" ]; then
  cd_lane_drain_limits_ok=1
fi
cd_lane_drain_capacity_ok=0
cd_lane_drain_labels_ok=0
if grep -Eq "^[[:space:]]+capacity:[[:space:]]*1[[:space:]]*$" /home/wooo/awoooi-cd-lane-drain/config.yaml 2>/dev/null; then
  cd_lane_drain_capacity_ok=1
fi
if grep -q "awoooi-ubuntu:docker://192.168.0.110:5000/awoooi/ci-runner:act-22.04" /home/wooo/awoooi-cd-lane-drain/config.yaml 2>/dev/null \
  && grep -q "awoooi-host:host" /home/wooo/awoooi-cd-lane-drain/config.yaml 2>/dev/null \
  && ! grep -Eq "^[[:space:]]+- \".*(ubuntu-latest|stockplatform|headless|playwright)" /home/wooo/awoooi-cd-lane-drain/config.yaml 2>/dev/null; then
  cd_lane_drain_labels_ok=1
fi
cd_lane_drain_binary_kind=$(file -b /home/wooo/awoooi-cd-lane-drain/awoooi_cd_lane_controlled 2>/dev/null || echo missing)
cd_lane_drain_binary_elf=0
echo "$cd_lane_drain_binary_kind" | grep -qi "ELF" && cd_lane_drain_binary_elf=1
cd_lane_drain_process_count=$(pgrep -f "^/home/wooo/awoooi-cd-lane-drain/awoooi_cd_lane_controlled" 2>/dev/null | wc -l | tr -d " ")
cd_lane_drain_ok=0
cd_lane_drain_mode=blocked
if [ "$cd_lane_drain_active" != "active" ] \
  && [ "$cd_lane_drain_binary_elf" = "0" ] \
  && [ "$cd_lane_drain_process_count" = "0" ] \
  && { [ "$cd_lane_drain_load" = "not-found" ] || { [ "$cd_lane_drain_load" = "masked" ] && [ "$cd_lane_drain_unitfile" = "masked" ]; }; }; then
  cd_lane_drain_ok=1
  cd_lane_drain_mode=failclosed
elif [ "$cd_lane_drain_active" = "active" ] \
  && [ "$cd_lane_drain_capacity_ok" = "1" ] \
  && [ "$cd_lane_drain_labels_ok" = "1" ] \
  && [ "$cd_lane_drain_binary_elf" = "1" ] \
  && [ "$cd_lane_drain_limits_ok" = "1" ]; then
  cd_lane_drain_ok=1
  cd_lane_drain_mode=controlled_open
fi
echo "CD_LANE_DRAIN_CONTROLLED mode=$cd_lane_drain_mode load=$cd_lane_drain_load unitfile=$cd_lane_drain_unitfile active=$cd_lane_drain_active capacity=$cd_lane_drain_capacity_ok labels=$cd_lane_drain_labels_ok binary_elf=$cd_lane_drain_binary_elf limits=$cd_lane_drain_limits_ok process_count=$cd_lane_drain_process_count ok=$cd_lane_drain_ok"
cd_lane_root_restore_left=unknown
if sudo -n true >/dev/null 2>&1; then
  cd_lane_root_restore_left=$(sudo -n find /root -maxdepth 1 -type d \( -name "awoooi-cd-lane-disabled-*" -o -name "awoooi-cd-lane-drain-disabled-*" \) -print 2>/dev/null | wc -l | tr -d " ")
fi
echo "CD_LANE_ROOT_RESTORE_SOURCES left=$cd_lane_root_restore_left"
cd_lane_guard_ok=0
if { [ "$cd_lane_ok" = "1" ] || [ "$cd_lane_drain_ok" = "1" ]; } && [ "$cd_lane_root_restore_left" = "0" ]; then
  cd_lane_guard_ok=1
fi
echo "CD_LANE_GUARDRAILS_OK $cd_lane_guard_ok"
[ "$cd_lane_guard_ok" = "1" ] || bad=1
direct_runner_count=$(pgrep -f "^/home/wooo/act-runner/act_runner|^/home/wooo/act-runner-controlled/act_runner|^/home/wooo/awoooi-controlled-runner/awoooi_controlled_runner" 2>/dev/null | wc -l | tr -d " ")
echo "RUNNER_DIRECT_PROCESS_COUNT $direct_runner_count"
[ "$direct_runner_count" = "0" ] || bad=1
for p in /home/wooo/act-runner/act_runner /home/wooo/act-runner/act_runner.real-20260628-runner-pressure-guard /home/wooo/act-runner-controlled/act_runner /home/wooo/awoooi-controlled-runner/awoooi_controlled_runner; do
  kind=$(file -b "$p" 2>/dev/null || echo missing)
  echo "RUNNER_FAILCLOSED_BINARY $p kind=$kind"
  echo "$kind" | grep -qi "ELF" && bad=1
done
for u in $(systemctl list-units "actions.runner.*" --all --no-legend --plain 2>/dev/null | awk "{print \$1}"); do
  load=$(systemctl show "$u" -p LoadState --value)
  unitfile=$(systemctl show "$u" -p UnitFileState --value)
  mainpid=$(systemctl show "$u" -p MainPID --value)
  watchdog=$(systemctl show "$u" -p WatchdogUSec --value)
  quota=$(systemctl show "$u" -p CPUQuotaPerSecUSec --value)
  memory=$(systemctl show "$u" -p MemoryMax --value)
  state=$(systemctl show "$u" -p ActiveState --value)
  action_ok=0
  action_mode=blocked
  if [ "$state" != "active" ] \
    && { [ "$load" = "masked" ] || [ "$load" = "not-found" ] || [ "$unitfile" = "masked" ] || [ "$unitfile" = "disabled" ]; } \
    && [ "${mainpid:-0}" = "0" ]; then
    action_ok=1
    action_mode=github_disabled
  fi
  echo "$u mode=$action_mode load=$load unitfile=$unitfile state=$state mainpid=$mainpid watchdog=$watchdog quota=$quota memory=$memory ok=$action_ok"
  [ "$action_ok" = "1" ] || bad=1
done
echo "BAD_RUNNER_GUARDRAILS $bad"
' 2>&1); then
    blocked "runner guardrail check unavailable"
    echo "$out"
    return
  fi
  echo "$out"
  grep -q "BAD_RUNNER_GUARDRAILS 0" <<<"$out" && ok "legacy runner fail-closed and controlled cd-lane guardrails complete" || blocked "legacy runner / controlled cd-lane guardrails incomplete"
}

check_job_containers() {
  section "active job container CPU"
  local out
  if ! out=$(ssh_cmd "wooo@192.168.0.110" '
names=$(docker ps --format "{{.Names}}" | grep -E "^(GITEA-ACTIONS-|awoooi-cd-)" || true)
if [ -z "$names" ]; then
  echo "NO_ACTIVE_JOB_CONTAINERS"
  exit 0
fi
for name in $names; do
  cpu=$(docker stats "$name" --no-stream --format "{{.CPUPerc}}" | tr -d "%" | awk "{printf \"%.6f\", \$1 / 100}")
  echo "JOB_CONTAINER $name cpu_cores=$cpu"
done
' 2>&1); then
    blocked "job container CPU check unavailable"
    echo "$out"
    return
  fi
  echo "$out"
  if grep -q "NO_ACTIVE_JOB_CONTAINERS" <<<"$out"; then
    ok "no active Gitea/CD job containers"
    return
  fi
  local bad_count
  bad_count=$(awk -v limit="$JOB_CONTAINER_CPU_LIMIT" -F'cpu_cores=' '/^JOB_CONTAINER / {if (($2 + 0) > limit) bad++} END {print bad+0}' <<<"$out")
  if [ "$bad_count" -eq 0 ]; then
    ok "active job containers are below ${JOB_CONTAINER_CPU_LIMIT} CPU cores"
  else
    blocked "$bad_count active job container(s) exceed ${JOB_CONTAINER_CPU_LIMIT} CPU cores"
  fi
}

check_high_load_services() {
  section "high-load service health"
  local out
  if out=$(ssh_cmd "ollama@192.168.0.188" '
echo "ollama-systemd $(systemctl is-active ollama 2>/dev/null || true)"
echo "ollama-api $(curl -s -o /dev/null -w "%{http_code}" --max-time 5 http://127.0.0.1:11434/api/tags || true)"
docker inspect -f "momo-scheduler {{.State.Status}} {{if .State.Health}}{{.State.Health.Status}}{{end}}" momo-scheduler 2>/dev/null || true
docker inspect -f "litellm {{.State.Status}} {{if .State.Health}}{{.State.Health.Status}}{{end}}" litellm 2>/dev/null || true
if ! docker inspect litellm >/dev/null 2>&1 && [ ! -d /opt/litellm ]; then
  echo "litellm not-provisioned"
fi
docker inspect -f "signoz-clickhouse {{.State.Status}} {{if .State.Health}}{{.State.Health.Status}}{{end}}" signoz-clickhouse 2>/dev/null || true
' 2>&1); then
    echo "$out"
    grep -q "ollama-systemd active" <<<"$out" && ok "188 Ollama systemd active" || blocked "188 Ollama systemd inactive"
    grep -q "ollama-api 200" <<<"$out" && ok "188 Ollama API reachable" || blocked "188 Ollama API not reachable"
    grep -q "momo-scheduler running healthy" <<<"$out" && ok "188 momo-scheduler healthy" || blocked "188 momo-scheduler not healthy"
    if grep -Eq "litellm running( |$)" <<<"$out"; then
      ok "188 litellm running"
    elif grep -q "litellm not-provisioned" <<<"$out"; then
      warn "188 litellm not provisioned; provider route/cost switch requires separate approval"
    else
      blocked "188 litellm not running"
    fi
    grep -q "signoz-clickhouse running healthy" <<<"$out" && ok "188 SignOz ClickHouse healthy" || warn "188 SignOz ClickHouse health not confirmed"
  else
    blocked "188 high-load service check unavailable"
    echo "$out"
  fi

  if out=$(ssh_cmd "wooo@192.168.0.110" '
docker inspect -f "sentry-clickhouse {{.State.Status}} {{if .State.Health}}{{.State.Health.Status}}{{end}}" sentry-self-hosted-clickhouse-1 2>/dev/null || true
docker inspect -f "sentry-kafka {{.State.Status}} {{if .State.Health}}{{.State.Health.Status}}{{end}}" sentry-self-hosted-kafka-1 2>/dev/null || true
docker ps --format "{{.Names}} {{.Status}}" | grep -E "sentry-self-hosted-(snuba|events|transactions|generic|metrics|subscription).*consumer" | head -20 || true
' 2>&1); then
    echo "$out"
    grep -q "sentry-clickhouse running healthy" <<<"$out" && ok "110 Sentry ClickHouse healthy" || blocked "110 Sentry ClickHouse not healthy"
    grep -q "sentry-kafka running healthy" <<<"$out" && ok "110 Sentry Kafka healthy" || blocked "110 Sentry Kafka not healthy"
    grep -q "Restarting" <<<"$out" && blocked "110 Sentry consumers include restarting containers" || ok "110 sampled Sentry consumers are not restarting"
  else
    blocked "110 high-load service check unavailable"
    echo "$out"
  fi
}

summary() {
  section "summary"
  echo "PASS=$PASS WARN=$WARN BLOCKED=$FAIL"
  if [ "$FAIL" -gt 0 ]; then
    echo "Result: HOLD_P3_RELEASE. Do not release runner/CD/crawlers/consumers further."
    return 1
  fi
  if [ "$WARN" -gt 0 ]; then
    echo "Result: P3_RELEASE_WITH_CAUTION. Proceed only with operator review."
    return 0
  fi
  echo "Result: P3_RELEASE_READY. Controlled high-load work release is allowed."
}

echo "AWOOOI P3 controlled release gate"
date '+%Y-%m-%d %H:%M:%S %Z'
echo "Limits: load5/core<=$LOAD5_PER_CORE_LIMIT load15/core<=$LOAD15_PER_CORE_LIMIT job_container_cpu<=$JOB_CONTAINER_CPU_LIMIT"

check_cold_start_gate
check_host_load "110" "wooo@192.168.0.110"
check_host_load "188" "ollama@192.168.0.188"
check_textfiles
check_storage_health
check_backup_health
check_runner_guardrails
check_job_containers
check_high_load_services
summary