Zigbee data request

[abe]

Hi, Im developing a Zigbee concentrator. I used nRF5_SDK_for_Thread_and_Zigbee_v3.1.0_c7c4730 ble_zigbee_template as start point of my project.
The light bulbs and switch control is working well, the first sample are on the field for test:
Now I'm approching to IAS zone sensors.
My question: how can I check if sensor is connected and working?
I see on the sniffer that battery device sends timed requests (see image above). Can I intercept with a callback this communications?
Many thanks for answer
Abele

[Edvin]

Hello,

Do you know what sort of message this is? Have you tried to bind the coordinator to this cluster, and to enable subscription on it?

Please see the multisensor example test description for how to bind and subscribe to a cluster.

Best regards,

Edvin

[abe]

Hello,
this is not a specific cluster (I used bind and subcribe procedures for cluster and attributes with other end nodes). My question is another. This is radio level protocol data sent from battery powered end node as a "keep alive" signal.
Alexa use this info to update his internal connected device list:
In other words, where on the radio level protocol I can intercept this info/request?
Abele

[Edvin]

Hello,

I haven't received any response on this internal ticket either [ref your other case]. I have pinged this one as well. Can you please upload the full sniffer trace that your screenshot is from so I can add it to the internal ticket? Just in case they ask for it, to save some time.

BR,

Edvin

[abe]

Hi Edvin, I'm sorry for delay, I was busy for other works...
Here attached the sniffer trace file. You can see the behavior as in the image starting from line 286
Many thanks, regards
Abele

DataRequest.pcapng

[abe]

Hello, any news about?

[Edvin]

Hello,

Reply from Zigbee team:

What about using Poll Control Cluster Check-In command instead? This seems as a bit better solution. Also I am pretty sure that we don't have some sort of data-request callback/handler exposed to the application (why should we?).

Data requests are sent from end device to the parent device only. If the coordinator is not within end device radio range it will not even hear the frame.
Check-In commands are sent to bound devices with configured interval and are received by the application using zcl_device_cb(), seems easier to check if the device is alive or not.

What about this solution?

[abe]

Hello, I will try the solution as you suggest. I will keep you updated.
Many thanks for help
Abele