fix(stats): flush throttled daily mileage projections

docs/ops/go-service-observability.md

@@ -150,6 +150,7 @@ systemctl start lingniu-go-capacity-check.service
 | `vehicle_stat_samples_total` | Mileage samples extracted and written by stat writer. Use this with `vehicle_stat_writes_total`: write `ok` only means the append call succeeded, while `status="written"` proves a sample reached `vehicle_daily_mileage_source`. |
 | `vehicle_stat_sources_total` | Stat writer source tracking results for `vehicle_data_source`. Labels: `topic`, `protocol`, `status`; status includes `attempted`, `written`, `skipped_throttled`, `skipped_missing_endpoint`. This intentionally avoids source IP labels to keep metrics low-cardinality. |
 | `vehicle_stat_projections_total` | Final daily mileage projection results from `vehicle_daily_mileage_source` to `vehicle_daily_mileage`. Labels: `topic`, `protocol`, `status`; status includes `attempted`, `written`, `skipped_throttled`. |
+| `vehicle_stat_pending_projections_total` / `vehicle_stat_pending_projection_entries` | Eventual projection tail flush. `attempted` and `written` should grow together; `error` indicates a pending final projection remains queued for retry. The entries gauge should stay bounded by active vehicle/protocol/day keys and drain after the projection interval. |
 | `vehicle_realtime_updates_total` | Redis/MySQL realtime projector updates. Labels: `topic`, `status`. |
 | `vehicle_realtime_kafka_messages_total` | Realtime projector Kafka message results by raw topic. Labels: `topic`, `status`; `invalid_json` messages are committed and isolated before Redis/MySQL projection. |
 | `vehicle_realtime_last_message_unix_seconds` | Last realtime Kafka message by topic/status. Labels: `topic`, `status`. |

docs/ops/vehicle-ingest-runbook.md

@@ -342,7 +342,8 @@ export MYSQL_DSN="$(sed -n 's/^MYSQL_DSN=//p' /opt/lingniu-go-native/env/base.en
 | `vehicle_stat_samples_total{status="written"}` / `recovered_stationary` | stat-writer `ok` 增长但 `written` 不增长 | fields topic 有消费但没有有效里程样本入库，按 `skipped_missing_vin`、`skipped_missing_mileage`、`skipped_non_positive_mileage`、`skipped_missing_source`、`skipped_same_mileage` 定位原因；如果主要是 `skipped_same_mileage`，通常表示上游总里程未变化，不等同故障。`recovered_stationary` 表示宇通明确静止帧已用同来源历史仪表值生成 `0 km`，应与 `skipped_missing_mileage` 联合观察。 |
 | `vehicle_stat_sources_total{status="written"}` / `skipped_throttled` | `vehicle_stat_samples_total{status="found"}` 增长但 source tracking 无增长 | 来源管理链路没有维护 `vehicle_data_source`。优先查 source endpoint 是否为空、`vehicle_data_source` 表结构、MySQL 写入错误和 `STATS_SOURCE_TOUCH_INTERVAL_SECONDS`。 |
 | `vehicle_stat_sources_total{status="skipped_missing_endpoint"}` | 单 topic 缺 endpoint / found 超过 `capacity-check -stat-source-max-missing-ratio` | fields envelope 缺 `source_endpoint`，无法区分平台来源和维护 source。优先查 Gateway 是否填充 TCP remote/MQTT endpoint，以及 NATS/Kafka bridge 是否保留 envelope 字段。 |
-| `vehicle_stat_projections_total{status="written"}` / `skipped_throttled` | source 候选持续写入但最终日里程更新频率低 | `vehicle_daily_mileage_source` 是每样本候选事实层，`vehicle_daily_mileage` 是最终投影层。`skipped_throttled` 增长通常表示受 `STATS_PROJECT_INTERVAL_SECONDS` 节流，不代表样本丢失；如需强一致核验可临时设投影间隔为 `0`。 |
+| `vehicle_stat_projections_total{status="written"}` / `skipped_throttled` | source 候选持续写入但最终日里程更新频率低 | `vehicle_daily_mileage_source` 是每样本候选事实层，`vehicle_daily_mileage` 是最终投影层。`skipped_throttled` 增长通常表示受 `STATS_PROJECT_INTERVAL_SECONDS` 节流，不代表样本丢失；被节流的末条样本会进入待投影队列，即使车辆停止上报也会在后续周期完成尾刷。如需强一致核验可临时设投影间隔为 `0`。 |
+| `vehicle_stat_pending_projections_total{status="error"}` / `vehicle_stat_pending_projection_entries` | 尾刷失败或待投影队列持续增长 | 先检查 MySQL 可写性、死锁和连接超时。失败项不会丢弃，会保留到下一个周期重试；健康状态下 `attempted` 与 `written` 应同步增长，entries 应保持在当日活跃车辆/协议键数量以内并周期性回落。 |
 | `vehicle_history_last_*_unix_seconds` / `vehicle_stat_last_*_unix_seconds` / `vehicle_realtime_last_*_unix_seconds` | 已出现过的成功 activity 超过 `capacity-check -last-activity-stale-seconds` 未刷新 | 下游消费、写入或 commit 某 topic 停止活动；结合 Kafka lag 和对应存储健康判断。 |
 | `vehicle_stat_project_interval_seconds` | 生产异常为 `0` 或被调得过小 | `vehicle_daily_mileage_source` 仍每样本更新，但 `vehicle_daily_mileage` 选举投影会被节流；过小会放大 MySQL 压力。 |
 | Kafka lag | 连续 5 分钟增长或 `> 10000` | 下游 consumer 或存储存在瓶颈。 |