Fleet telematics data: the platform we build for it
Fleet telematics data arrives as several streams at different rates, and each has its own storage and retention needs. Agnotic Technologies builds the ingestion, storage and ELD data handling to your fleet's requirements.
Last reviewed by the Agnotic Technologies engineering team
What does a fleet telematics data platform need to handle?
A fleet telematics data platform must ingest GPS pings, engine readings, driving events, ELD records and camera clip references from every vehicle, absorb the busiest second of traffic, and keep each stream for its own retention period. Agnotic Technologies builds ingestion, time-series storage and ELD data handling so each stream lands in the right place.
Which telematics data problems do fleets run into?
Telematics projects rarely fail on maps or dashboards. They struggle with volume, mixed streams and retention.
The ping interval drives everything
Pings per vehicle equal running seconds divided by the seconds between pings. Halve the interval and the ping count doubles, along with ingestion load and storage for that stream.
Streams with different triggers
GPS pings follow a timer, driving events follow thresholds and ELD records follow duty status changes. Treating them as one feed makes sizing and querying harder.
ELD records follow their own rules
Hours of service rules set what each ELD record holds, how long it is kept and that a backup copy sits on a separate device. Mixing these records into analytics history breaks those rules.
Ingestion sized for the average falls behind
A broker or queue sized for the daily average backs up when many vehicles run at once, which delays live maps and alerts.
Hot storage fills with old history
Live maps and alerts query recent months, while audits and claims read older data. Without lifecycle rules, the hot store keeps everything at the most expensive tier.
Video mixed into the telematics store
Dashcam clips belong in object storage with a reference row in the telematics store. Storing them alongside pings slows every query.
What do we build for fleet telematics data?
Agnotic Technologies builds the telematics platform to your requirements: your devices, your streams and the features your operations team needs.
Device ingestion layer
An MQTT broker or message queue sized for peak messages per second, with a decoder per device type that maps each payload to a common message format.
Time-series storage
A store partitioned by time and vehicle, with partition size set from expected messages per day and indexes chosen for the queries your screens run.
Hot and archive lifecycle rules
Jobs that move older history from the hot store to compressed files on object storage after the months you choose, with a query path for audits.
ELD record store
Separate tables for ELD records with their own retention, a backup copy on a separate device and access rules for roadside inspection and audit.
Event and alert processing
Rules for harsh driving events, geofence entries and exits, and engine diagnostic codes, with alerts routed to dispatch and maintenance.
Camera clip index
A reference row per clip that links each event to its video in object storage, so reviewers find the right footage from the event.
Ingestion monitoring
Dashboards and alerts on message lag, decoder errors and vehicles that stop reporting, so gaps are found before a report depends on them.
Which data types does a telematics device send?
Each stream has its own trigger, size and retention reason, so each one is planned separately.
| Data type | What triggers it | What it holds | Why you keep it |
|---|---|---|---|
| GPS ping | A timer while the engine runs | Position, speed, heading and any sensor values sent with it | Live map, trip history, geofence checks |
| Engine reading | A timer or a change in a monitored value | Engine hours, odometer, fuel level, diagnostic codes | Maintenance planning and fault alerts |
| Driving event | A threshold crossed, such as harsh braking | Event type, time, location and the readings around it | Driver coaching and incident review |
| ELD record | A duty status change, an hourly intermediate recording, a login or a certification | Date, time, location, engine hours, miles and driver, vehicle and carrier IDs | Hours of service compliance and roadside inspection |
| Camera clip reference | An event or a request from the office | Clip ID, time, vehicle and storage location; the video sits in object storage | Finding the video for an event |
How does the ping interval change the volume?
Pings per vehicle per day equal running hours times 3,600, divided by the seconds between pings. Halve the interval and the ping count doubles, along with ingestion load and storage for that stream.
That makes the interval the setting to decide first. Dispatch screens and geofence alerts may need short intervals, while trip history and mileage reports can work from fewer points.
Some devices also send extra pings on turns or speed changes. If yours do, measure pings per running hour from a week of real data before you plan capacity.
Why are ELD records a separate stream?
Electronic logging device records exist for hours of service rules, not for fleet analytics, and the rules set what each record holds and how long it is kept.
An ELD records a set list of data elements at every duty status change and makes an intermediate recording when the vehicle has been moving for an hour without one. Motor carriers keep records of duty status and supporting documents for at least 6 months, with a backup copy on a separate device.
Because that retention period differs from your telematics history, ELD records belong in their own tables with their own retention, apart from the years of ping history you keep.
Hot, archive and backup tiers
Hot time-series store
Recent months that live maps, alerts and dashboards query. Plan it from recent message volume, then add headroom for indexes in the store you choose.
Archive
Older history in compressed files on object storage, read for audits, claims and model training.
Replicas and backups
Copies multiply both tiers. Count a second region or a backup system once each.
How does Agnotic build a telematics platform?
Agree the streams, devices and ping intervals the platform must support, and the features each stream feeds.
Build the ingestion layer, such as an MQTT broker or a message queue, for peak messages per second, not the daily average.
Partition the time-series store by time and vehicle, and set lifecycle rules that move data from hot to archive.
Keep the ELD stream in its own tables with its own retention and access rules.
Connect the data to the screens that use it. Our fleet telematics platform page shows where each stream lands, and the fleet management software page covers the work the data supports.
Which rules set ELD record content and retention?
Federal Motor Carrier Safety Administration rules in 49 CFR Part 395, read in the eCFR on 21 Oct 2026.
Under 49 CFR 395.26(b), an ELD automatically records date, time, vehicle location, engine hours, vehicle miles and driver, vehicle and motor carrier identification data.
Source: 49 CFR 395.26, ELD data automatically recorded (eCFR)Under 49 CFR 395.26(d), when a vehicle is in motion and there has been no duty status change or intermediate recording in the previous hour, the ELD automatically makes an intermediate recording.
Source: 49 CFR 395.26(d), Intermediate recording (eCFR)49 CFR 395.8(k)(1) requires a motor carrier to retain records of duty status and supporting documents for each driver for not less than 6 months from the date of receipt.
Source: 49 CFR 395.8(k), Retention of records of duty status (eCFR)49 CFR 395.22(i)(1) requires a motor carrier to retain a back-up copy of ELD records for 6 months on a device separate from the one that stores the original data.
Source: 49 CFR 395.22(i), Record backup and security (eCFR)
Trusted by global innovators
Where should you go next?
- Fleet telematics platformDevice ingestion, live maps and alerts.
- Fleet management softwareVehicles, maintenance, drivers and compliance in one system.
- Transportation and logistics analyticsKPIs and reporting on fleet and shipment data.
- Fleet management glossary entryWhat fleet management software covers, in plain terms.