An MQTT # Subscription Can Miss $ Topics: Test the Filter Boundary
A broad MQTT subscription receives ordinary telemetry but misses an expected system topic. That does not automatically mean the broker is offline or the message was lost. MQTT 3.1.1 treats topic names beginning with $ differently when a subscription filter begins with a wildcard. Check the matching boundary before widening access or changing the publisher.

Key takeaways |
A leading # or + filter does not match a topic beginning with $. An explicit authorised filter may be needed, but the broker must actually expose the topic. Subscription matching and access permission are separate checks. |
1. Write down the exact filter and topic
Record the topic name as published and the filter as submitted by the receiving client. Preserve case, separators and any leading character. Copying a displayed label from an application is not enough: it may be an alias rather than the protocol topic. Keep credentials out of the diagnostic record.
The ProSight Software platform overview is a starting point for a monitoring integration discussion. Confirm the deployed broker and supported interfaces with the project team. This article does not imply that a particular system topic or broker administration view is exposed by ProSight Software.
Separate three questions: did the publisher send the expected topic, would the filter match it, and is the subscriber authorised to receive it? Evidence for one does not settle the other two. A successful connection alone establishes none of them.
2. Apply the leading-dollar rule
OASIS MQTT 3.1.1 section 4.7.2 excludes dollar-prefixed topic names from filters that start with # or +. Consequently, treating # as literally every possible topic can produce an incomplete diagnostic capture. Use the exact protocol version and broker documentation when designing a test.
The standard permits brokers to use leading-dollar names for other purposes. It does not promise that every broker implements a particular $SYS hierarchy. An explicit subscription cannot create a topic or force the broker to expose operational information it does not publish.
For a documented and authorised system namespace, consider the narrow explicit filter specified by the broker owner. Do not grant a diagnostic user broad cross-customer or administrative access merely because an initial wildcard test returned no data.

3. Build a small matching matrix
In an isolated fixture, define one ordinary topic and one documented dollar-prefixed test or system topic. Record which filters should match each according to MQTT 3.1.1. Keep topic matching separate from the server’s publication policy and the client’s access rights.
Start by proving the ordinary topic with its exact subscription. Then test the broader filter and compare the result. Test the documented dollar namespace explicitly only where the server owner permits it. A real system namespace may require a controlled test environment rather than arbitrary publication.
Capture the subscription request and acknowledgement alongside received messages. If the subscription is rejected, stop treating the result as a wildcard-matching experiment. Resolve the permission or supported-filter issue through the normal access process.
Diagnostic checks
Result | Interpretation to test |
Exact ordinary topic works; # works | Ordinary route is demonstrated |
$ topic works explicitly; # misses it | Leading-dollar matching boundary |
Explicit filter also fails | Publication, permission or topic availability |
4. Diagnose an empty result without guessing
An empty result can mean that no message was published during the observation window, that the filter did not match, that access was refused or that the application discarded the message after receipt. Put timestamps beside each observation and identify the last stage with positive evidence.
Check the documented publication interval and whether the expected state is retained. Avoid changing production publication intervals just to make a screenshot look populated. A bounded test should establish the route and mapping without unnecessary load or configuration churn.
If an authorised explicit filter receives the expected message while # does not, record the leading-dollar distinction in the handover. If neither receives it, continue with broker documentation, publication evidence and access policy rather than assuming the protocol rule explains every failure.
5. Keep the operational subscription narrow
Broad diagnostic subscriptions can reveal more operational data than a normal application needs. After testing, restore the approved least-privilege filters and verify that required asset topics remain visible. Keep the accepted filter set with the equipment and integration configuration.
Retest after namespace, broker or authorisation changes. Include an ordinary telemetry case and a documented system-topic case where relevant. The aim is an explainable boundary, not a permanent subscription to all available traffic.
Recommended equipment / related solutions
Use the ProSight Software platform overview to discuss monitoring and reporting scope. For a tailored interface and acceptance plan, review ProSight Custom. Confirm the broker, deployment and supported integration before implementation.
Related guide: MQTT Topic and Payload Design for Industrial IoT Dashboards. It covers the wider naming and payload context; this article addresses a specific protocol boundary.
Frequently asked question
Does $SYS exist on every MQTT broker?
No. The standard allows dollar-prefixed server topics but does not define a universal $SYS inventory. Consult the exact broker documentation and access policy.
Technical reference
OASIS MQTT 3.1.1 specification. Protocol scope in this guide is MQTT 3.1.1; project test and handover steps are engineering recommendations.
Need help with your application?
Discuss your ProSight Software integration requirement. Send the documented broker/client versions, a sanitised example and the required acceptance outcome.

.png)
Comments