Thread is a mesh, not a star network — every powered router negotiates with neighbors before sleepy sensors get a reliable path home. When leader-router handshakes time out, the mesh cannot elect stable parents. Accessories show Connecting, automations skip triggers, and rebooting one bulb mysteriously fixes the whole floor for an hour.
Handshake timeouts sit one layer below what most apps show you. Wi-Fi can look perfect while Thread routers fail to agree on who leads the local partition.
Symptoms that point to handshake failure
Devices that worked yesterday suddenly show No Response in clusters — not randomly across the house. Border routers rebooted overnight (automatic updates, power blips) and never fully reformed the mesh.
In manufacturer diagnostics, parents flip between None and a router that is physically too far away. Matter commissioning stalls at the last step while the accessory LED shows joined-to-Thread behavior.
Multiple border routers from different vendors online at once increases handshake churn after any single hub restarts.
Baseline checks before deep surgery
Confirm mains-powered Thread routers — Apple TV, HomePod, Nest hub, plugged-in Eve extenders, OTBR hosts — are awake and on the same LAN as your phone. VPN on the phone does not break Thread locally, but it makes app diagnostics misleading.
Update firmware on border routers before accessories. A router on older Thread stack timing may timeout talking to newly updated sleepy children.
Reduce 2.4 GHz congestion: microwave ovens, older cordless phones, and crowded Zigbee channels all raise handshake retry rates even when RSSI looks fine.
Practical recovery steps
Reboot in mesh order: gateway router first, wait five minutes, then border routers from the center of the home outward. Do not power-cycle every end device at once.
Temporarily reduce router count: if you have three ecosystems advertising border routing, power down all but the one that owns most devices for a twelve-hour test. Handshakes often complete when only one vendor leads.
Force a local re-parent: move one failing sleepy device within three meters of a known-good border router for thirty minutes. If handshakes succeed there, add another powered router midway on that floor — not more resets.
Remove stale entries: ghost accessories in Apple Home or Google Home keep radios polling dead routes. Delete unresponsive duplicates before re-adding hardware.
Ethernet backhaul: border routers on weak mesh Wi-Fi backhaul drop leader elections under load. Wire at least the primary hub.
When timeouts indicate a bigger problem
Handshake storms that return daily on a wired, simple network may need vendor support with border router logs. Do not edit Thread network keys manually from forum snippets.
If a specific border router overheats or reboots independently, repair hardware before chasing mesh parameters.
Confirm handshakes are stable
Pick two sleepy devices at opposite ends of the home. Trigger them on a schedule every thirty minutes for a day. Leader changes should happen without No Response gaps longer than a few minutes after hub reboot tests.