feat(msg): 完整实现 msg 消息服务

包含 channels/preferences/templates/grpc/kafka/outbox/push/redis 等完整实现

services/msg/src/shared/observability/metrics.ts

@@ -4,25 +4,234 @@ const registry = new promClient.Registry();
 // 修复：在本地 registry 上设置默认标签（原代码误用全局 register）
 registry.setDefaultLabels({ service: "msg" });
 
+// ============================================================
+// HTTP 指标（02-architecture-design.md §6.4）
+// ============================================================
+
 registry.registerMetric(
   new promClient.Counter({
-    name: "msg_requests_total",
-    help: "Total number of msg requests",
-    labelNames: ["method", "endpoint", "status"],
+    name: "msg_http_requests_total",
+    help: "Total number of msg HTTP requests",
+    labelNames: ["method", "route", "status_code"],
   }),
 );
 
 registry.registerMetric(
   new promClient.Histogram({
-    name: "msg_request_duration_seconds",
-    help: "Msg request duration in seconds",
-    labelNames: ["method", "endpoint"],
+    name: "msg_http_request_duration_seconds",
+    help: "Msg HTTP request duration in seconds",
+    labelNames: ["method", "route"],
     buckets: [0.01, 0.05, 0.1, 0.3, 0.5, 1, 3, 5],
   }),
 );
 
+// ============================================================
+// gRPC 指标
+// ============================================================
+
+registry.registerMetric(
+  new promClient.Counter({
+    name: "msg_grpc_requests_total",
+    help: "Total number of msg gRPC requests",
+    labelNames: ["rpc_method", "status"],
+  }),
+);
+
+// ============================================================
+// 业务指标：通知发送
+// ============================================================
+
+registry.registerMetric(
+  new promClient.Counter({
+    name: "msg_notification_sent_total",
+    help: "Total number of notifications sent",
+    labelNames: ["type", "channel", "status"],
+  }),
+);
+
+registry.registerMetric(
+  new promClient.Histogram({
+    name: "msg_notification_send_duration_seconds",
+    help: "Notification send duration in seconds",
+    labelNames: ["type", "channel"],
+    buckets: [0.01, 0.05, 0.1, 0.3, 0.5, 1, 3, 5],
+  }),
+);
+
+registry.registerMetric(
+  new promClient.Counter({
+    name: "msg_notification_failed_total",
+    help: "Total number of notification send failures",
+    labelNames: ["type", "channel", "error_code"],
+  }),
+);
+
+// ============================================================
+// 业务指标：通知已读
+// ============================================================
+
+registry.registerMetric(
+  new promClient.Counter({
+    name: "msg_notification_read_total",
+    help: "Total number of notifications marked as read",
+    labelNames: ["type"],
+  }),
+);
+
+// ============================================================
+// 业务指标：渠道分发
+// ============================================================
+
+registry.registerMetric(
+  new promClient.Histogram({
+    name: "msg_channel_dispatch_duration_seconds",
+    help: "Channel dispatch duration in seconds",
+    labelNames: ["channel"],
+    buckets: [0.01, 0.05, 0.1, 0.3, 0.5, 1, 3, 5],
+  }),
+);
+
+// ============================================================
+// Kafka 消费指标
+// ============================================================
+
+registry.registerMetric(
+  new promClient.Gauge({
+    name: "msg_kafka_consumer_lag",
+    help: "Kafka consumer lag (offset difference)",
+    labelNames: ["topic"],
+  }),
+);
+
+registry.registerMetric(
+  new promClient.Counter({
+    name: "msg_kafka_consumer_processed_total",
+    help: "Total number of Kafka messages processed",
+    labelNames: ["topic", "status"],
+  }),
+);
+
+// ============================================================
+// 幂等去重指标
+// ============================================================
+
+registry.registerMetric(
+  new promClient.Counter({
+    name: "msg_idempotent_duplicate_total",
+    help: "Total number of idempotent duplicate hits",
+    labelNames: ["topic"],
+  }),
+);
+
+// ============================================================
+// Outbox 指标
+// ============================================================
+
+registry.registerMetric(
+  new promClient.Gauge({
+    name: "msg_outbox_pending_count",
+    help: "Number of pending outbox events",
+  }),
+);
+
+// ============================================================
+// Push Gateway 指标
+// ============================================================
+
+registry.registerMetric(
+  new promClient.Counter({
+    name: "msg_push_gateway_call_total",
+    help: "Total number of Push Gateway calls",
+    labelNames: ["status"],
+  }),
+);
+
+registry.registerMetric(
+  new promClient.Histogram({
+    name: "msg_push_gateway_call_duration_seconds",
+    help: "Push Gateway call duration in seconds",
+    buckets: [0.01, 0.05, 0.1, 0.3, 0.5, 1, 3, 5],
+  }),
+);
+
+// ============================================================
+// 业务状态指标
+// ============================================================
+
+registry.registerMetric(
+  new promClient.Gauge({
+    name: "msg_unread_count",
+    help: "User unread notification count (sampled)",
+    labelNames: ["user_id"],
+  }),
+);
+
 // 自动收集 Node.js 进程级指标（CPU/内存/事件循环/GC等）
 // 这些指标无需业务代码埋点，prom-client 自动采集
 promClient.collectDefaultMetrics({ register: registry });
 
-export { registry as metricsRegistry };
+// ============================================================
+// 便捷访问器：业务代码通过命名导出获取已注册的 Metric 实例
+// ============================================================
+
+export const metricsRegistry = registry;
+
+export const httpRequestsTotal = registry.getSingleMetric(
+  "msg_http_requests_total",
+) as promClient.Counter<string>;
+
+export const httpRequestDurationSeconds = registry.getSingleMetric(
+  "msg_http_request_duration_seconds",
+) as promClient.Histogram<string>;
+
+export const grpcRequestsTotal = registry.getSingleMetric(
+  "msg_grpc_requests_total",
+) as promClient.Counter<string>;
+
+export const notificationSentTotal = registry.getSingleMetric(
+  "msg_notification_sent_total",
+) as promClient.Counter<string>;
+
+export const notificationSendDurationSeconds = registry.getSingleMetric(
+  "msg_notification_send_duration_seconds",
+) as promClient.Histogram<string>;
+
+export const notificationFailedTotal = registry.getSingleMetric(
+  "msg_notification_failed_total",
+) as promClient.Counter<string>;
+
+export const notificationReadTotal = registry.getSingleMetric(
+  "msg_notification_read_total",
+) as promClient.Counter<string>;
+
+export const channelDispatchDurationSeconds = registry.getSingleMetric(
+  "msg_channel_dispatch_duration_seconds",
+) as promClient.Histogram<string>;
+
+export const kafkaConsumerLag = registry.getSingleMetric(
+  "msg_kafka_consumer_lag",
+) as promClient.Gauge<string>;
+
+export const kafkaConsumerProcessedTotal = registry.getSingleMetric(
+  "msg_kafka_consumer_processed_total",
+) as promClient.Counter<string>;
+
+export const idempotentDuplicateTotal = registry.getSingleMetric(
+  "msg_idempotent_duplicate_total",
+) as promClient.Counter<string>;
+
+export const outboxPendingCount = registry.getSingleMetric(
+  "msg_outbox_pending_count",
+) as promClient.Gauge<string>;
+
+export const pushGatewayCallTotal = registry.getSingleMetric(
+  "msg_push_gateway_call_total",
+) as promClient.Counter<string>;
+
+export const pushGatewayCallDurationSeconds = registry.getSingleMetric(
+  "msg_push_gateway_call_duration_seconds",
+) as promClient.Histogram<string>;
+
+export const unreadCount = registry.getSingleMetric(
+  "msg_unread_count",
+) as promClient.Gauge<string>;