#!/usr/bin/env bash
set -Eeuo pipefail

# 采集主机、Docker、PostgreSQL、MySQL 诊断信息。
# 用法:
#   bash deploy/collect-db-diagnostics.sh
#   bash deploy/collect-db-diagnostics.sh /tmp/diagnostics.txt
#
# 可覆盖的环境变量:
#   PG_SERVICE=postgres PG_USER=postgres PG_DB=order_site
#   MYSQL_CONTAINER=douyu-login-db MYSQL_USER=root MYSQL_PASSWORD=...

PROJECT_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
OUTPUT_FILE="${1:-${PROJECT_ROOT}/order-site-diagnostics-$(date +%Y%m%d-%H%M%S).txt}"
PG_SERVICE="${PG_SERVICE:-postgres}"
PG_USER="${PG_USER:-postgres}"
PG_DB="${PG_DB:-order_site}"
MYSQL_CONTAINER="${MYSQL_CONTAINER:-douyu-login-db}"
MYSQL_USER="${MYSQL_USER:-root}"
MYSQL_PASSWORD="${MYSQL_PASSWORD:-}"

mkdir -p "$(dirname "${OUTPUT_FILE}")"
: >"${OUTPUT_FILE}"
cd "${PROJECT_ROOT}"

FAILURES=0

write_header() {
  printf '\n===== %s =====\n' "$1" >>"${OUTPUT_FILE}"
}

run_cmd() {
  local label="$1"
  shift
  write_header "${label}"
  {
    printf '$'
    printf ' %q' "$@"
    printf '\n'
  } >>"${OUTPUT_FILE}"
  set +e
  "$@" >>"${OUTPUT_FILE}" 2>&1
  local status=$?
  set -e
  printf '[exit=%s]\n' "${status}" >>"${OUTPUT_FILE}"
  if [[ ${status} -ne 0 ]]; then
    FAILURES=$((FAILURES + 1))
  fi
}

run_pg_sql() {
  local label="$1"
  local sql="$2"
  write_header "PostgreSQL: ${label}"
  printf 'SQL:\n%s\n' "${sql}" >>"${OUTPUT_FILE}"
  set +e
  docker compose exec -T "${PG_SERVICE}" \
    psql -X -v ON_ERROR_STOP=1 -U "${PG_USER}" -d "${PG_DB}" \
    -P pager=off -P footer=on -c "${sql}" >>"${OUTPUT_FILE}" 2>&1
  local status=$?
  set -e
  printf '[exit=%s]\n' "${status}" >>"${OUTPUT_FILE}"
  if [[ ${status} -ne 0 ]]; then
    FAILURES=$((FAILURES + 1))
  fi
}

run_mysql_sql() {
  local label="$1"
  local sql="$2"
  write_header "MySQL: ${label}"
  printf 'SQL:\n%s\n' "${sql}" >>"${OUTPUT_FILE}"
  set +e
  if [[ -n "${MYSQL_PASSWORD}" ]]; then
    docker exec -e "MYSQL_PWD=${MYSQL_PASSWORD}" "${MYSQL_CONTAINER}" \
      mysql --batch --raw --show-warnings -u"${MYSQL_USER}" -e "${sql}" \
      >>"${OUTPUT_FILE}" 2>&1
  else
    docker exec "${MYSQL_CONTAINER}" \
      mysql --batch --raw --show-warnings -u"${MYSQL_USER}" -e "${sql}" \
      >>"${OUTPUT_FILE}" 2>&1
  fi
  local status=$?
  set -e
  printf '[exit=%s]\n' "${status}" >>"${OUTPUT_FILE}"
  if [[ ${status} -ne 0 ]]; then
    FAILURES=$((FAILURES + 1))
  fi
}

run_cmd "metadata" date -Is
run_cmd "host" uname -a
run_cmd "uptime" uptime
run_cmd "cpu" nproc
run_cmd "memory" free -h
run_cmd "vmstat (5 samples)" vmstat 1 5
run_cmd "filesystem" df -h
run_cmd "docker ps" docker ps --format 'table {{.Names}}\t{{.Status}}\t{{.RunningFor}}'
run_cmd "docker stats" docker stats --no-stream
run_cmd "docker system df" docker system df
run_cmd "docker disk usage" du -xhd1 /var/lib/docker /var/log

run_pg_sql "server settings" \
  "SELECT version(), current_database(), current_user; SELECT name, setting, unit FROM pg_settings WHERE name IN ('max_connections','shared_buffers','work_mem','maintenance_work_mem','effective_cache_size','track_io_timing','log_min_duration_statement','autovacuum','autovacuum_max_workers') ORDER BY name;"
run_pg_sql "activity summary" \
  "SELECT COALESCE(state,'<null>') AS state, COALESCE(wait_event_type,'<none>') AS wait_event_type, COALESCE(wait_event,'<none>') AS wait_event, count(*) AS connections FROM pg_stat_activity GROUP BY 1,2,3 ORDER BY connections DESC;"
run_pg_sql "active queries" \
  "SELECT pid, application_name, client_addr, state, wait_event_type, wait_event, clock_timestamp()-query_start AS duration, left(regexp_replace(query, '[[:space:]]+', ' ', 'g'), 1200) AS query FROM pg_stat_activity WHERE state <> 'idle' AND pid <> pg_backend_pid() ORDER BY query_start;"
run_pg_sql "blocking and locks" \
  "SELECT blocked.pid AS blocked_pid, clock_timestamp()-blocked.query_start AS blocked_for, blocking.pid AS blocking_pid, clock_timestamp()-blocking.query_start AS blocking_for, left(regexp_replace(blocked.query, '[[:space:]]+', ' ', 'g'), 500) AS blocked_query, left(regexp_replace(blocking.query, '[[:space:]]+', ' ', 'g'), 500) AS blocking_query FROM pg_stat_activity blocked JOIN LATERAL unnest(pg_blocking_pids(blocked.pid)) AS b(pid) ON true JOIN pg_stat_activity blocking ON blocking.pid=b.pid ORDER BY blocked.query_start;"
run_pg_sql "top pg_stat_statements" \
  "SELECT calls, round(total_exec_time::numeric,2) AS total_ms, round(mean_exec_time::numeric,2) AS mean_ms, rows, left(regexp_replace(query, '[[:space:]]+', ' ', 'g'), 1000) AS query FROM pg_stat_statements ORDER BY total_exec_time DESC LIMIT 30;"
run_pg_sql "largest tables" \
  "SELECT schemaname, relname, pg_size_pretty(pg_total_relation_size(relid)) AS total_size, pg_size_pretty(pg_relation_size(relid)) AS table_size, pg_size_pretty(pg_indexes_size(relid)) AS index_size, n_live_tup, n_dead_tup, last_autovacuum, last_autoanalyze FROM pg_stat_user_tables ORDER BY pg_total_relation_size(relid) DESC LIMIT 30;"
run_pg_sql "dead tuple and vacuum health" \
  "SELECT schemaname, relname, n_live_tup, n_dead_tup, CASE WHEN n_live_tup > 0 THEN round(100.0*n_dead_tup/(n_live_tup+n_dead_tup),2) ELSE 0 END AS dead_pct, last_autovacuum, last_vacuum, last_autoanalyze, last_analyze, autovacuum_count, autoanalyze_count FROM pg_stat_user_tables ORDER BY n_dead_tup DESC LIMIT 30;"
run_pg_sql "worker query indexes" \
  "SELECT tablename, indexname, indexdef FROM pg_indexes WHERE tablename IN ('work_orders','work_order_shares','work_order_events','worker_order_cancel_requests','worker_order_feedbacks') ORDER BY tablename, indexname;"
run_pg_sql "retention table counts" \
  "SELECT c.relname AS table_name, c.reltuples::bigint AS estimated_rows, pg_size_pretty(pg_total_relation_size(c.oid)) AS total_size FROM pg_class c JOIN pg_namespace n ON n.oid=c.relnamespace WHERE n.nspname='public' AND c.relname IN ('orders','kuaishou_industry_vouchers','task_events','webhook_events','admin_audit_logs','work_order_events') ORDER BY pg_total_relation_size(c.oid) DESC;"

run_mysql_sql "server status" \
  "SELECT VERSION() AS version, @@hostname AS hostname, @@max_connections AS max_connections, @@innodb_buffer_pool_size AS innodb_buffer_pool_size, @@innodb_flush_log_at_trx_commit AS flush_log_at_trx_commit, @@sync_binlog AS sync_binlog;"
run_mysql_sql "process list" "SHOW FULL PROCESSLIST;"
run_mysql_sql "innodb status" "SHOW ENGINE INNODB STATUS;"
run_mysql_sql "largest tables" \
  "SELECT table_schema, table_name, engine, table_rows, ROUND((data_length+index_length)/1024/1024,2) AS total_mb, ROUND(data_length/1024/1024,2) AS data_mb, ROUND(index_length/1024/1024,2) AS index_mb FROM information_schema.tables WHERE table_schema NOT IN ('information_schema','performance_schema','mysql','sys') ORDER BY data_length+index_length DESC LIMIT 30;"
run_mysql_sql "innodb counters" \
  "SHOW GLOBAL STATUS WHERE Variable_name IN ('Threads_connected','Threads_running','Created_tmp_tables','Created_tmp_disk_tables','Innodb_buffer_pool_reads','Innodb_buffer_pool_read_requests','Innodb_buffer_pool_pages_dirty','Innodb_log_waits','Innodb_row_lock_waits','Innodb_row_lock_time','Innodb_data_reads','Innodb_data_writes','Innodb_os_log_fsyncs');"

write_header "next step"
cat >>"${OUTPUT_FILE}" <<'EOF'
EXPLAIN (ANALYZE, BUFFERS) requires the real worker ID parameter. The active-query
SQL above is intentionally captured without replaying it, so this collection does
not modify application data. For a slow query, run EXPLAIN only in a maintenance
window and replace $1 with a representative worker ID:

  EXPLAIN (ANALYZE, BUFFERS, VERBOSE) <captured SQL with $1 replaced by a value>;
EOF

printf '诊断采集完成: %s (failed checks: %s)\n' "${OUTPUT_FILE}" "${FAILURES}"
printf '把这个文件完整发回即可。\n'