ops: add production observability and recovery

infra/scripts/restore.sh

@@ -0,0 +1,47 @@
+#!/bin/sh
+set -eu
+
+if [ "$#" -ne 1 ] || [ "${RESTORE_CONFIRM:-}" != "restore-ai-ticket-platform" ]; then
+  echo "用法：RESTORE_CONFIRM=restore-ai-ticket-platform $0 <backup-directory>" >&2
+  echo "恢复会替换当前 PostgreSQL 数据库和 MinIO 数据。" >&2
+  exit 2
+fi
+
+SCRIPT_DIR=$(CDPATH= cd -- "$(dirname -- "$0")" && pwd)
+INFRA_DIR=$(CDPATH= cd -- "$SCRIPT_DIR/.." && pwd)
+COMPOSE_FILE=${COMPOSE_FILE:-"$INFRA_DIR/docker-compose.prod.yml"}
+TARGET=$(CDPATH= cd -- "$1" && pwd)
+
+"$SCRIPT_DIR/verify-backup.sh" "$TARGET"
+
+POSTGRES_DB=${POSTGRES_DB:-ai_ticket_platform}
+POSTGRES_USER=${POSTGRES_USER:-ai_ticket}
+BACKUP_HELPER_IMAGE=${BACKUP_HELPER_IMAGE:-alpine:3.21}
+
+if [ "${RESTORE_SKIP_APPLICATION_SERVICES:-false}" != "true" ]; then
+  docker compose -f "$COMPOSE_FILE" stop backend worker beat
+fi
+MINIO_CONTAINER=$(docker compose -f "$COMPOSE_FILE" ps -q minio)
+if [ -z "$MINIO_CONTAINER" ]; then
+  echo "未找到 MinIO 容器" >&2
+  exit 3
+fi
+docker compose -f "$COMPOSE_FILE" stop minio
+docker compose -f "$COMPOSE_FILE" exec -T postgres \
+  dropdb -U "$POSTGRES_USER" --force "$POSTGRES_DB"
+docker compose -f "$COMPOSE_FILE" exec -T postgres \
+  createdb -U "$POSTGRES_USER" "$POSTGRES_DB"
+docker compose -f "$COMPOSE_FILE" exec -T postgres \
+  pg_restore -U "$POSTGRES_USER" -d "$POSTGRES_DB" --no-owner --exit-on-error \
+  < "$TARGET/postgres.dump"
+docker run --rm \
+  --volumes-from "$MINIO_CONTAINER" \
+  -v "$TARGET:/backup:ro" \
+  "$BACKUP_HELPER_IMAGE" \
+  sh -c 'find /data -mindepth 1 -maxdepth 1 -exec rm -rf -- {} + && tar -C /data -xzf /backup/minio-data.tar.gz'
+docker compose -f "$COMPOSE_FILE" up -d --wait minio
+if [ "${RESTORE_SKIP_APPLICATION_SERVICES:-false}" != "true" ]; then
+  docker compose -f "$COMPOSE_FILE" start backend worker beat
+fi
+
+echo "恢复完成。请立即检查 /api/system/health/ready 并执行核心业务冒烟测试。"