Discovery

Discovery helps you build your Unified Namespace from your existing MQTT traffic. HiveMQ detects MQTT topics that do not match any namespace node and lists them as Ungoverned topics. You then review each ungoverned topic and either approve it into the namespace or ignore it.

Discovery complements the top-down approach in Define Namespace, where you design the structure first and align your data to it. Discovery suits projects with existing infrastructure that already produces valuable data, where rapidly reflecting the current state of operations is a priority. You can combine both approaches: start with Discovery to accelerate onboarding, then refine the model top-down for governance and long-term manageability.

Discovery works with the following glossary terms:

  • Ungoverned Data: Data that is not yet captured or managed in the namespace.

  • Governed Data: Data that is modeled and managed in the namespace.

  • Ignored Topic: A topic that is explicitly skipped during namespace operations.

View Ungoverned MQTT Traffic

To view the MQTT traffic on a connected broker, open the Discovery tab in Contextualize. If your project has no broker that supports Discovery, the tab shows an introduction and an option to connect a broker.

Discovery introduction before a broker is connected

Select a Broker

  1. Above the traffic list, click the broker selector to expand the broker table.

  2. Optional: In the search field, enter the name of the broker that you want to view.

  3. In the table, select a connected broker, such as HQ Broker.

Broker selector with search and page navigation

You can select only one broker at a time. The table also lists offline brokers, but you cannot select them. The table shows up to 10 brokers per page. If more pages exist, use the first, previous, next, or last page controls in the footer. You can also enter a page number in Go to page. The Deviations tab uses the same broker selector.

Inspect the Traffic List

Ungoverned MQTT traffic in the Discovery view

The Ungoverned MQTT Traffic panel organizes detected traffic into two sub-tabs:

  • Discovered Topics: Topics the platform has detected but that are not yet part of the namespace. These are candidates for approval.

  • Ignored Topics: Topics you have explicitly chosen to skip during namespace operations.

To narrow the list, use the topic filter input, which supports the MQTT wildcards + and #. The Wrap long Topic toggle controls how the panel displays full topic strings. A count (for example, Showing 1 - 4 of 4 Discovered Topics) shows how many topics match. Each topic has a checkbox so you can select one or more for review.

To get the latest traffic from the selected broker, click the refresh control above the list.

If the selected broker has no discovered topics, the panel shows an empty result. Check that the broker receives traffic on topics outside the namespace. Also check that the topic filter matches those topics. Traffic for inactive namespace nodes does not appear as discovered traffic.

During a request, the panel shows a progress indicator. If the request takes longer, the panel shows an additional message about data from the broker. An error appears separately from an empty result. If a request fails or times out, check the broker connection and click the refresh control to retry.

Review a Discovered Topic

Click a topic in the Discovered Topics list to open its detail view, titled Ungoverned Topic. The detail view shows the topic (for example, acme/berlin/assembly/line-1/conveyor-belt/temperature) and provides everything you need to decide whether to bring it into the namespace.

Reviewing the details of a discovered topic

The General tab presents the following information:

Field Description

Namespace

The breadcrumb showing where the topic sits in the namespace hierarchy, such as acme > berlin > assembly > line-1 > conveyor-belt > temperature. A copy button is provided.

MQTT Topic

The full MQTT topic the data flows on. A copy button is provided.

Suggested Data Type

The data type the platform infers from the observed payloads, such as Decimal. You can adjust this before you approve it.

Suggested Serialization Format

The serialization format the platform infers, such as UTF8. You can adjust this before you approve it.

Data Sample

A sample payload observed on the topic, such as 20.0, to support your decision.

Message Count

The number of messages observed on the topic, along with the approximate rate in messages per second (MPS), such as 35 at ~0.34 MPS.

First seen at / Last seen at

Timestamps for when the platform first and most recently observed the topic.

When a discovered topic carries a complex payload, such as a JSON object, a Fields tab appears alongside General. The Fields tab lists the individual elements in the payload that are stored in the namespace as fields, together with their auto-selected types.

You cannot edit field types during approval, but you can modify them individually after you add the topic to the namespace.

Approve a Topic Into the Namespace

When you approve a discovered topic, HiveMQ creates an inactive node for the topic in your namespace. The node uses the data type, serialization format, and topic hierarchy that Discovery suggests.

  • If your knowledge of the data source differs from the platform’s inference, adjust the Suggested Data Type and Suggested Serialization Format before approval.

  • Click Approve into Namespace to create the node.

  • If the topic requires parent nodes that do not yet exist, the platform adds them to the namespace automatically. The platform highlights any new parent nodes, so you can see what it created.

All settings remain editable after approval. You can refine the node at any time once it is part of the namespace.
An approved node enters the namespace inactive. When you activate it, HiveMQ stores its history, runs its computations, and counts it toward your Data Intelligence bill. For details, see Node Activation.

Approve or Ignore Topics in Bulk

When a broker exposes many topics, you can act on them together instead of reviewing each one. In the Discovered Topics list, use the checkboxes to select individual topics, or the header checkbox to select every topic the list currently holds.

Bulk selection of discovered topics for approval

When you select one or more topics, the toolbar shows the number of selected topics (for example, 12 Topics Selected) along with two actions:

  • Approve Selection into Namespace: Adds every selected topic to the namespace in one step. HiveMQ creates an inactive node for each selected topic. Each node uses the data type, serialization format, and topic hierarchy that Discovery suggests. You can add multiple topics at the same time. You do not need to review each topic separately.

  • Ignore Selected Traffic: Moves every selected topic to the Ignored Topics sub-tab in one step.

Discovery loads at most 1000 topics at a time. When a broker holds more, approve or ignore the loaded topics to reach the next batch.

Bulk approval is the fastest way to populate a namespace from a broker that already carries structured traffic. You can refine any individual node afterward.

Ignore a Topic

If a discovered topic does not belong in the namespace, click Ignore Topic. The topic moves to the Ignored Topics sub-tab, and namespace operations skip it.

To ignore several topics at once, use Ignore Selected Traffic, as described in Approve or Ignore Topics in Bulk.

Review and Restore Ignored Topics

Open the Ignored Topics sub-tab to see every topic you ignored on the selected brokers. The list shows each topic with the data type the platform inferred for it, and an Unignore button.

Click Unignore to bring a topic back into review. The platform can discover the topic again, so it returns to Discovered Topics when the broker next reports it.

Unignore acts on one topic at a time. The Ignored Topics list offers no bulk selection.

Next Steps

After you bring discovered data into your namespace, you can refine and extend the structure: