Files
lingniu-vehicle-ingest/docs/superpowers/plans/2026-06-30-kafka-streaming-mileage.md
2026-07-01 01:21:09 +08:00

41 lines
1.4 KiB
Markdown

# JT808 Kafka Streaming Mileage Plan
## Current Goal
Consume JT808 Kafka location events, keep rolling daily state, and write the derived daily mileage into the common vehicle-stat metric repository.
## Current Design
- Source topic: `vehicle.event.jt808.v1`
- Runtime app: `vehicle-analytics-app`
- State: Redis or in-memory `Jt808MileageStateStore`
- Metric output: `VehicleStatRepository.saveDailyStat(...)`
- Default metric storage: `VEHICLE_STAT_FILE_PATH/daily-stats.tsv`
- Date boundary: `Asia/Shanghai`
JT808 daily mileage is calculated only from the GPS total mileage reported by JT808 location additional information:
```text
daily_mileage_km = last_total_mileage_km - first_total_mileage_km
```
The stream saves a metric only when at least two ordered location points for the same vehicle and local day contain valid `total_mileage_km`. Negative differences are treated as invalid and are not written.
## Runtime Properties
```text
KAFKA_TOPIC_JT808_EVENT=vehicle.event.jt808.v1
VEHICLE_STAT_ENABLED=true
VEHICLE_STAT_JT808_MILEAGE_ENABLED=true
VEHICLE_STAT_JT808_STATE_STORE=redis
VEHICLE_STAT_FILE_PATH=/data/vehicle-stat
REDIS_HOST=<host>
REDIS_PORT=6379
REDIS_DATABASE=50
REDIS_PASSWORD=<password>
```
## Notes
Do not create or write a protocol-specific JT808 daily-mileage table. Do not add GPS segment distance, speed integral, or jump-filter state back into this path unless the mileage definition changes again.