feat: build vehicle data platform and production pipeline
This commit is contained in:
@@ -69,7 +69,7 @@
|
||||
|
||||
- `history.Writer.AppendAllBatch` 按 raw child table 和 location child table 生成多行 `INSERT ... VALUES (...),(...)`。
|
||||
- oversized payload chunks 仍写入 `raw_frame_payload_chunks`,同样按 child table 批量写。
|
||||
- `history-writer` Kafka 消费默认按 `HISTORY_BATCH_SIZE=200`、`HISTORY_BATCH_WAIT_MS=100` 收集消息。
|
||||
- `history-writer` 默认启动 `HISTORY_WORKERS=3` 个同组消费者,每个 worker 按 `HISTORY_BATCH_SIZE=200`、`HISTORY_BATCH_WAIT_MS=20` 收集其已分配分区的消息。
|
||||
- TDengine batch 成功后才批量提交 Kafka messages;batch 失败不提交 offset,让 Kafka 保留可重放语义。
|
||||
- 保留 `processHistoryMessage` 和 `AppendAll` 单条路径,便于回退和测试。
|
||||
|
||||
|
||||
Reference in New Issue
Block a user