The driver-status and resilience stores had no removal path, so a deleted
driver stayed on /hosts as a frozen row until restart. Both stores gain
RetainOnly, which /hosts calls after each successful config load (never on a
failed one).
The stores and InProcessBroadcaster now raise their events one subscriber at a
time, so a throwing handler cannot starve later subscribers or escape onto the
bridge actor. The alert/script-log bridges publish inside their try, and the
resilience bridge catches a store failure instead of restarting.
Refs #745
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
CallTimeoutSeconds was bound, defaulted and documented but consumed
nowhere. TelemetryDialSupervisor now arms a per-generation deadline on
every dial: a dial that has delivered no event by the deadline is
cancelled and treated as a transport failure (reconnect with a failure
streak). A live node replays its snapshots on subscribe, so this bites
only a dial that never connects (a black-holed host, where the socket
connect has no timeout). An established stream is deliberately not
bounded — the wire has no heartbeat, keepalive pings own its liveness.
Review ref: 02/U2.
Refs #731
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Joseph Doherty