top of page

MQTT Topic and Payload Design for Industrial IoT Dashboards

Oct 31
4 min read

MQTT is a widely used publish-subscribe messaging protocol for connected devices.

Illustrative industrial MQTT telemetry architecture with gateway, broker and dashboards; not a live ProSight screen.

Original ProSight Software concept illustration; not a live software screenshot or customer installation.

What you will learn

  • Use stable asset identifiers in topics.

  • Include measurement time, engineering units and quality.

  • Handle duplicate and delayed messages explicitly.

  • Test the complete sensor-to-dashboard path.

How MQTT topics, message payloads, timestamps and quality flags work together to deliver reliable industrial telemetry.

It can carry telemetry from gateways and edge applications to software systems, but the protocol alone does not define how equipment names, units, timestamps or data quality should be represented.

Those decisions belong to the solution architecture. A consistent topic and payload design makes dashboards, alarm rules and integrations easier to maintain as a project grows.

1. Understand the three main roles

A publisher sends messages to a broker under a topic. A subscriber requests messages matching a topic filter. The broker distributes messages according to the configured subscriptions and permissions.

MQTT does not, by itself, transform raw measurements into an industrial historian or guarantee that every downstream dashboard has stored every reading.

An edge gateway might publish measurements collected from Modbus RTU instruments, while a monitoring application subscribes to the relevant telemetry.

The gateway must still handle polling, register interpretation, engineering-unit conversion, connectivity and local buffering according to the actual device and application requirements.

2. Build a topic hierarchy around stable identifiers

A useful topic scheme is predictable and not tied to a person's display preferences. For example, an illustrative structure could be sites/site-01/assets/pump-03/telemetry.

Use stable asset identifiers rather than changing descriptive names. Separate production and test environments, and document whether a topic represents an individual measurement, an equipment state or a grouped payload.

MQTT topics are case-sensitive. A leading slash creates a different topic from the same text without that slash. Topic wildcards are subscription filters, not characters to embed casually in published topic names.

Keep the naming rules simple enough that field technicians can diagnose messages without guessing.

3. Include enough context in the payload

A telemetry payload should allow the receiving system to interpret a reading correctly. Consider a timestamp, measurement key, numeric value, engineering unit and quality indicator.

A sample JSON message might describe measured pressure in kilopascals alongside the time it was captured and a status such as valid or invalid. The exact schema should be agreed before devices are deployed.


Field

Purpose

asset_id

Stable equipment identifier, such as pump-03

sample_time

Time the measurement was captured, with explicit timezone

value + unit

Numeric measurement and its agreed engineering unit

quality

Whether the measurement is valid, invalid or stale


Illustrative payload fields only. Agree the actual schema with the receiving application before deployment.

Distinguish measurement time from broker receipt time. During an outage, a gateway may buffer data and forward it later.

If the cloud system stamps only arrival time, a delayed historical reading could appear to be a new event.

Use a consistent timestamp format, preferably with an explicit UTC offset or UTC representation, and document clock synchronisation expectations.

Four telemetry contract elements: topic, payload, ingestion and dashboard.

Illustrative design guide. Confirm the actual settings and supported functions for your application.

4. Handle units and data quality explicitly

A number without a unit is ambiguous. A pressure reading could be in bar, kilopascals or psi; a temperature value could be Celsius or Fahrenheit.

Agree on canonical units and conversions, and do not mix units silently across sensors. Store scaling and calibration context in the asset configuration where appropriate.

Quality flags help differentiate a genuine zero reading from a communication error, stale value or out-of-range sensor.

This is especially important when Modbus register failures, gateway restarts or disconnected sensors occur. A dashboard should not display a last-known value as though it were current without indicating its age.

5. Choose delivery behaviour to match the use case

MQTT quality-of-service levels describe message delivery between MQTT clients and the broker; they do not automatically guarantee that a business application has processed and permanently stored a measurement.

Duplicate delivery can occur in some operating conditions. Design downstream ingestion so repeated messages do not corrupt totals or create duplicate events.

Retained messages can be useful for the latest known state, but they are not a historical database. Last Will messages can help indicate an unexpected client disconnect, subject to broker and client configuration.

For historical trends and audit requirements, specify an appropriate time-series storage and retention design separately.

6. Secure the connection and topic access

Use transport encryption where supported, unique device credentials, and least-privilege permissions for publishing and subscribing. Avoid one shared account with access to every customer or site.

Rotate compromised credentials and keep broker, gateway and certificate lifecycle responsibilities documented.

Industrial networks should be segmented according to risk. Avoid exposing controllers or gateways directly to the public internet simply to make dashboard integration convenient.

Where remote access is needed, assess secure connectivity options and the consequences of losing the network.

7. Validate with a commissioning checklist

Before handover, verify the exact topic, asset identifier, payload schema, engineering units, timestamp interpretation, message frequency, quality states and access controls.

Test broker disconnection and reconnection, duplicate messages, out-of-order historical data and invalid payloads. Confirm the dashboard shows an understandable error rather than silently accepting corrupted data.

8. Connect MQTT to a broader monitoring architecture

MQTT is one part of an industrial monitoring system. Sensor compatibility, gateway configuration, broker operations, ingestion rules, data retention, visualisations and alerting all need to work together.

ProSense can assist with measurement and communications selection.

Confirm the features of any proposed ProSight deployment against the currently available platform rather than assuming planned capabilities are already released.

Practical next step

Define one representative asset, its telemetry schema and its failure modes. Test the entire path from sensor to dashboard with realistic network interruptions.

Once the conventions are proven, reuse them consistently across sites and equipment families.

Related ProSight guides and resources

Comments


prosight-software-logo (1).png
bottom of page