400 trucks, one live map.
A regional carrier replaced a server in the office closet with an AWS telemetry pipeline that ingests a GPS ping every eight seconds from 400 vehicles and drives the dispatch board in real time.
- Client
- Meridian Freight Lines
- Industry
- Logistics & fleet
- Location
- Red Deer, Alberta
- Timeline
- 18 weeks
- Team
- 3 engineers
Dispatch
LiveFleet Telemetry & Dispatch Cloud
Telemetry arrived through a single on-premise server that lost power twice a quarter. Dispatch worked from a whiteboard and drivers called in their positions.
Idling was invisible: no one could tell a truck waiting at a dock from one parked for three hours.
- AWS IoT
- Lambda
- TimescaleDB
- Grafana
What it looks like in use.
Idle time per vehicle
BeforeAfterFleet status
4 rowsFour decisions that made it work.
-
01
Decouple ingestion from reading
Pings land in a queue and are written by a time-series writer; the dispatch board reads from a cache, so a stalled dashboard never drops data.
-
02
Design for a dead zone
Vehicles buffer up to six hours of pings locally and backfill on reconnect, so mountain routes do not leave holes in the map.
-
03
Alert on behaviour, not noise
Idle over 12 minutes, geofence exits and harsh braking trigger a dispatch notification — with a mute per vehicle.
-
04
Migrate without a night shift
Both pipelines ran in parallel for five weeks; the cutover happened on a Tuesday morning with the old server switched off that afternoon.
“Dispatch used to be a whiteboard and a phone. Now the board is right, and it is right at 4am.”