An ops dashboard that runs on one Lambda
Replaced a suite of disconnected spreadsheets with a single internal tool — load status, driver assignments, and route exceptions in one view, updated in real time.
- Internal tool
- Lambda
- DynamoDB
- Real-time
- Operations
- Three spreadsheets → one dashboard
- Data sources consolidated
- Real-time via WebSocket
- Update latency
- Scales to zero between shifts
- Runtime cost
The single-page app and the Lambda API deploy from one pipeline through dev → staging → prod, with a manual approval before prod. Because the whole stack is infrastructure-as-code, a new environment for a new depot is one command — not a week of clicking.
Grove's dispatch team tracked load status in one spreadsheet, driver assignments in another, and route exceptions in a third. Reconciling the three was a manual task done at shift handover — errors surfaced late, and supervisors had no live view of the floor during a shift.
We modelled the domain as a stream of load and driver state changes, landing each event in a DynamoDB table with a time-bucketed sort key. A Lambda-backed WebSocket API pushes updates to the dashboard as they happen. The frontend is a single-page React app that renders the current state of every active load and driver without a page refresh. The whole stack is serverless — it scales to zero between shifts and to the full dispatch team in seconds.
Supervisors now have a live floor view for the duration of a shift. Exceptions surface immediately rather than at handover. The three spreadsheets were retired, and because the runtime is serverless, the tool costs near nothing when the depot is closed.
Let’s write the next system into being
Tell us what you’re building. We’ll reply within two business days with a frank read on scope, shape, and whether we’re the right studio for it.