Real-Time Fleet Monitoring Dashboard
Helios Robotics
Helios runs autonomous units across warehouses and outdoor yards. Operators were stitching together CSV exports and a 30-second polling UI — too slow to catch a unit drifting off-path. They needed a live picture of the fleet, plus a way to rewind an incident without a data-science ticket.
01 — The brief
The challenge
Helios's ops team was monitoring a fleet of 400+ autonomous units through a patchwork of spreadsheets and a legacy polling-based UI that refreshed every 30 seconds — far too slow to catch developing incidents before they became costly downtime.
02 — What we built
The solution
We built a real-time dashboard on WebSockets streaming telemetry directly from the fleet's edge gateways into a TimescaleDB time-series store, with configurable anomaly-detection rules and full historical playback for post-incident review.
03 — Method
How we got there
- 01Designed a WebSocket ingestion layer handling 12k events/second
- 02Modeled telemetry in TimescaleDB with continuous aggregates for fast queries
- 03Built configurable rule-based anomaly alerts with Slack/PagerDuty integration
- 04Shipped a historical playback mode for post-incident analysis
04 — Proof
What changed
- Reduced incident response time by 61%
- Handles 12k telemetry events/second
- Deployed across 3 regional operations centers
“We came to them with a fleet of robots and a spreadsheet. We left with a real-time dashboard our ops team actually loves using every day.”
05 — Frames
Inside the work
More Web Development
Ready to build something great?
Tell us about your project and we'll get back to you within one business day with next steps.