• Ultraloq Bolt Fingerprint lock (battery, Matter over Thread), firmware 02.37.0028. Multi-admin: Apple Home, SmartThings, Flic Hub.
• Thread border routers: two HomePod minis and an Apple TV 4K (27.2); two SmartThings Hub v3 (000.062.00016, separate Thread network, no devices).
• Measured with a Hub SDK module: Identify (IdentifyTime 0), reachable from getNodes() and nodeUpdate, connected from matter.subscribe().

I've see that a command that times out drops the node: sendCommand calls back after 70 to 175s (50: CHIP Error 0x00000032: Timeout where recorded), and the node goes unreachable in the same second. Eight times today. With nothing sent, the node dropped 10 minutes after the last answered command. After a drop, sendCommand fails in about 50ms with 72: node not connected, for minutes to hours. Today, answered commands took 44 to 112s. Ping to the lock through each border router works and takes about 2.7s. The node reported reachable and connected for 14 minutes, then the first command failed in 153ms with error 72 and then the flags go false. A hub restart reconnects the node within about 20s, but the first command after 7 of 8 restarts today failed and dropped it again. Last night, a test command running every 3 minutes all answered for 1.5h, until an automatic router reboot, and then connected never returned, though commands succeeded for another 3h. Today, most commands timed out, at intervals of 3, 7 and 15 minutes. Throughout all the issues via Flic, Apple Home and SmartThings continue control the lock, and all border routers announce its Thread prefix. Rebooting Thread hubs, battery pulls and re-adding the lock to the Flic Hub did not help.

Is this a known issue? Are there any logs I can collect or more details I can provide?