Multi-Node Integration Tests for Distributed Brain Simulation

Purpose: provide comprehensive integration tests for multi-node communication, error recovery, and performance characteristics of the distributed brain simulation.

Test structure - Unit: component behavior - Integration: multi-node workflows (PFC ↔ Vision, node discovery) - System: end-to-end scenarios and performance

Example test contents (summary) - PFC ↔ Vision communication tests (mocked or real Zenoh) - Node discovery and heartbeat/cleanup tests - Error recovery (retry/backoff) for ZenohCommunicator - Latency and throughput measurements (P95/P99) - Raft consensus leader election and log replication checks

Notes - Many tests guard imports so they can run in environments where optional dependencies (Zenoh, Raft) are not installed; they fall back to mocks. - Integration tests that require real Zenoh / distributed runners should be executed in an integration environment with daemons available.

If you want, I can extract selected test skeletons into a runnable integration test job that uses Docker Compose to bring up a local Zenoh daemon and three mock nodes.