URGENT: Event Grid MQTT in ZA NORTH no delivering messages

Hi @Dustyn Lightfoot

We are also experiencing the same issue with Azure Event Grid MQTT Broker today.

Our environment is showing the exact same behavior:

• MQTT clients connect successfully
• SUBACK and PUBACK are returned successfully
• Published messages appear as accepted
• However, messages are not being delivered to subscribers

When message delivery to our API stopped working properly, our devices automatically started retrying and resending data more frequently because acknowledgements and expected processing flows were no longer occurring correctly. As a consequence, the published message rate increased significantly after the incident started.

We have thousands of production devices impacted, affecting real users and customers here in Brazil.

What is especially concerning is that older MQTT sessions seem to continue working partially, while newer client connections are unable to properly exchange messages. We also observed dropped sessions and inconsistent message routing behavior.

This strongly appears to be a broker-side routing or delivery issue inside Azure Event Grid MQTT infrastructure, not a client-side problem.

Additionally, no configuration changes were made to our Event Grid MQTT infrastructure, permissions, topic spaces, client groups, or applications before the issue started. Our platform has been running stable in production for more than one year with the same configuration.

We urgently need confirmation from the Azure Event Grid team whether there is an ongoing incident, regional degradation, or recent platform change affecting MQTT delivery.

Environment:

• Azure Event Grid Namespace
• MQTT v3.1.1
• Production and development environments impacted
• Brazil-based IoT deployment with thousands of connected devices

Hi Siddhesh Desai,

Is the any ETA on when this will be resolved?

It is a production issue and is causing a lot of problems.

Kind Regards

Dustyn

Update from our side:

Yesterday, by the end of the afternoon, the messages that had not been delivered started being gradually delivered until the MQTT broker delivery functionality was fully restored.

No configuration changes, deployments, permission updates, topic space modifications, or infrastructure changes were made on our side during this period.

From our observations, Azure Event Grid MQTT Broker appeared to be unable to deliver messages in real time for several hours and eventually queued the messages internally until delivery slowly resumed.

At this moment, the Event Grid MQTT functionality appears to be operating normally again in our environment.

This reinforces our understanding that the issue was platform-side and not related to client configuration or application logic.