- Application Guide
- FlowFuse Guide
- Overview
- Foundations
- App delivery methods
- Hardware apps
- Software apps
- Data plane
- Architectures
- IT architectures
- OT architectures
- IIoT architectures
- Worked examples
- OEE, end to end
- Node-RED Guide
- Overview
- Foundations
- Patterns
- Design patterns
- Handling data
- Good form
- Worked examples
- OEE - Edge Aggregator
- OEE - Central Dashboard
OEE, end to end
The use case
Track OEE across three production lines. OEE (Overall Equipment Effectiveness) tells you how much good product a line makes versus its full potential — one live number per line, plus history for trends.
What "good" looks like — every line shows a live OEE figure the team trusts, and any line's OEE can be charted back over weeks to spot drift.
It's not one app. OEE takes two apps — one at the edge, one in the cloud — joined by two shared services (a broker and a history store).
The apps
1 · OEE - Edge Aggregator
What it does — Reads the machine signals on a line and publishes the line's state.
- Delivery method — whole app: promoted as a snapshot through a pipeline to a Remote Instance on every line.
- App pattern — hardware app · Configurable App: the same build everywhere, with each line's PLC tag names loaded as per-site config.
- Runs on — a Remote Instance, one per line.
- Why this shape — the same flows run on every line, right next to the equipment, and keep working if the link drops — but each line's PLC tags differ, so the values are configured per install. Build it once; roll the same version to the whole fleet and point each at its own tags.
- See the flow — how OEE - Edge Aggregator is built in Node-RED →
2 · OEE - Central Dashboard
What it does — Subscribes to the line states, computes availability, performance and quality, and presents the live OEE dashboard.
- Delivery method — whole app: promoted dev → staging → prod as one versioned build.
- App pattern — software app · Data-Driven App.
- Runs on — a Hosted Instance in the cloud.
- Why this shape — it's a user-facing app driven by live data, with no hardware of its own; one instance serves the whole plant.
- See the flow — how OEE - Central Dashboard is built in Node-RED →
Shared services
- Team Broker (MQTT · UNS) — carries line state from edge to cloud. The edge publishes to a topic; the dashboard subscribes. Neither references the other.
- External time-series DB (Timescale / QuestDB) — FlowFuse has no built-in time-series store, so history goes to an external DB over the Postgres wire protocol — a second egress, because the data is a timestamped stream.
The full architecture
The architecture, in one sentence — OEE is the Edge Aggregator (a hardware Configurable App, on a Remote Instance per line) publishing machine state over the Team Broker to the Central Dashboard (a software Data-Driven App, on a Hosted Instance), which computes and displays OEE and writes history to an external time-series DB.