Hello all friends,
Recently, many users have reported the issues of notifications delay. I would like to explain some of the reasons and the measures taken by Aqara.
-
All notifications have experienced delays: notifications from all devices in the Aqara Home app (sensor triggered events, camera detection events, system messages, etc.)delay, which is usually caused by message accumulation on cloud servers. In the past week, we found this situation happened on European servers, and Aqara has now expanded and upgraded the message processing capabilities of European servers. We believe that this phenomenon has been alleviated.
-
Only camera’s message delay: The notifications push mechanism of Aqara camera is as follows:
- Messages that are immediately pushed after triggering, including ringing the doorbell, forcibly removing the doorbell detected, etc. These events are sent by specific sensors of the device, and the type and triggering reason of the events are clear, so they are pushed immediately after they occur. When you still receive delayed messages, the reason must be network delay or issues processed by the aforementioned servers.
- The camera’s local AI recognizes messages, such as motion detection, person detection, etc. After these events are triggered, the camera first records a video of 6-30 seconds. When there are multiple event types in the video(such as when person detection triggers camera recording, but vehicles or pets appear during the recording process), the messages need to be summarized and displayed. Therefore, the delay time for such messages is fixed and not too long.
- The cloud AI recognition message of the camera usually has a specific home guarding icon on the app page (I have shown the cloud AI event in red circles in the figure below), and requires payment to continue activation. Their detection logic is to first trigger recording through motion detection, and then send the recorded video to the cloud server for analysis. After the analysis is completed, the results are pushed. Therefore, the delay fluctuation of the classification message is very large and random, completely dependent on the time consumption of the entire process and the processing speed of the server, we have recently increased the AI processing capability of our servers to alleviate the bottleneck of analysis. Moreover, if this type of detection is configured with automation control, it usually affects the duration of automation execution (as it can only be determined as an event after the AI analysis of the cloud server is completed)
When you notice a message delay in the app, please first confirm which type of problem it is based on the above principle. For the message push logic of the cameras mentioned in the second point, if you have good modification suggestions, please feel free to discuss them in the comment section or feedback directly to me.

