Shore Scale
Work / Traffic violation processing
200x
more events processed

From 2,000 to 400,000 traffic events an hour

Sensys Gatso GroupSensys GatsoSmart mobility · GovTech
Results
Throughput
2,000
→ 400,000
events/hour
Camera sites
56
→ 500+
IP cameras plus roadside sensors
Scaling
manual
→ autoscaled
on Kafka consumer lag
Human review
every case
→ edge cases only
clear-cut cases are processed automatically
The challenge

Every case may end up in court, so duplicates and weak evidence were never acceptable.

Hundreds of cameras and sensors produced a flood of raw events that had to become legally valid violation cases. Throughput was capped at 2,000 events an hour.

Architecture
100%
Switch layers on and off above, or hover a line in “What we built” to light up its layer.
What we built

From raw frames to court-ready cases.

  • 01 · INGESTION

    Ingestion layer: edge processing nodes and a redundant IoT gateway (MQTT → HTTP), with per-device mTLS.

  • 02 · STREAMING

    Streaming layer: a 6-node Kafka cluster with schema enforcement (Avro, Schema Registry) and S3 archival.

  • 03 · PROCESSING

    Processing layer: Go microservices for frame analysis, plate recognition, vehicle lookup, duplicate detection and evidence bundling.

  • 04 · WORKFLOW · REVIEW

    Workflow and review: confidence scoring sends uncertain cases to manual review and confident ones straight to case management.

  • 05 · INTEGRATIONS

    Integrations: national vehicle registry, postal notices, court filing and payments.

Stack
GogRPCKafkaAvroMQTTKubernetes · EKS, HPAPostgreSQLTimescaleDBRedisMinIOElasticsearchArgoCDPrometheusGrafanaJaeger

Is your pipeline drowning in events?

Talk to an engineer →
Next case
weather.com · Media & data
Kubernetes cache operator
94%
cache hit ratio