I have seen it more or less in the last few weeks that some of my Thread devices become unavailable and stay unavailable until I restart the Matter Server. In the attached log at 05:35:19.640 some of my devices started to become unavailable but recovered itself. Only nodes 10, 16 & 27 stayed unavailable until I restarted the Matter Server at 12:15:09.383. All devices are paired to Apple Border Routers (2x ATV, 2x HomePod mini) and shared to HA. The nodes being offline in HA worked in Apple Home at the same time.
#Some of my Matter over Thread devices become and stay unavailable until a Mater Server restart
quaint wedge
quaint wedge
@short blade or @silver surge, do you have an idea why some nodes do not recover itself and a restart of the Matter Server is necessary?
silver surge
Yeah from how I understand the issue is that some Apple Thread Border Routers don't route properly to devices. The routes pushed via IPv6 are all treated equal, so our Matter Server picks one at random. Restarting the Matter Server shuffles selection, likely selecting one that works. Afaik mostly the WiFi BRs are problematic. Try to shutdown the two HomePod minis for a while and see if that fixes the problem.
They work in Apple Home because Apple Home controller is always on the same device, likely an Apple TV in your case. So they don't route traffic through one of the problematic HomePods
timber sparrow
Resurrecting an ancient thread - is this still true, wrt to Apple devices not routing stuff? And if so, is there any appetite for adding some logic along the lines of 'ping known devices using each available route and select the one with the greatest percentage of good responses'? Asking before I've spent any time reading code, in case this has already been solved.