KIP-217 — Expose a timeout to allow an expired ZK session to be re-created

Discarded Kafka 1.1 Broker

Proposed adding a `zookeeper.reconnect.timeout.ms` config to bound how long a broker retries ZooKeeper reconnection before giving up and exiting, to allow Kubernetes/orchestrators to detect the failure and restart the pod. The existing behavior retries ZK session re-creation indefinitely with no way to signal failure to an external health checker. (Cancelled — resolved by doing infinite retry instead.)

Details