A border router only tracks so many Thread children at once. When the child table is full, new sensors cannot attach and existing ones may shed parents during reboots. This hits dense installs — lots of Eve, Aqara, and Nanoleaf gear on one Apple TV or Nest hub — not necessarily a broken primary router.
Recognize table pressure
Commissioning fails only for new devices while old ones still respond. Drops follow hub firmware updates or adding a dozen accessories in one afternoon. OTBR logs may mention child table limits if you run Home Assistant.
Relief that works
Add another mains-powered Thread router mid-home so children distribute across parents. Remove ghost accessories from Home apps — they still consume table entries on some stacks. Stage pairing: ten devices today, ten tomorrow. Update border router firmware; limits and eviction behavior change between releases. As last resort, factory reset straggler devices one at a time near a secondary router, not mass hub resets.
Verify
After changes, commission one new sleepy sensor at the network edge. It should join within minutes and survive overnight without No Response.
Official sources
- Thread Group — router roles and capacity
- OpenThread border router — child table behavior
- Apple Home hubs — placement guidance